research papers warehouse

The Effect of Warehousing Management on Warehouse Performance

102 Pages Posted: 19 Jan 2022

Ararsa Buzu

Jimma University (JU), College of Business and Economics

Date Written: July 20, 2021

In today’s challenging and competitive world, success can be determined by whether a warehouse management is productive and effective enough to meet the expectations of customers. The purpose of this study was examining the effect of warehousing management on warehouse performance in the case of Modjo dry port concerning the five main warehousing activities (receiving, put-away, storage, order-picking, and shipping). Both primary (questionnaires and interviews) and secondary sources of data were used. To achieve the objectives of this study, an explanatory and descriptive research design was used, and this study also applies a mixed research approach. Stratified simple random sampling was used to select the respondents for the study and, accordingly, one hundred one (101) sample sizes were taken for the study. The descriptive and inferential statistical tools such as; mean, standard deviation, percentage, correlation and multiple regressions were used to analyze collected data with the aid of IBM SPSS statistics version 20. The descriptive analysis shows that there is lack of space for loading and unloading items, lack of shelves, pallets and racks; poor well established put away process for received items, poor tight control the storage areas, high warehousing cost, and high inventory cost. The multiple regression analyses reveals that receiving, storage, put away, order picking and shipping significantly influence warehouse performance of the organization. Hence, organizations are expected to enhance their warehousing management so as to gain better warehouse performance. Furthermore, it’s advisable to the organization to improve the warehouse performance making use of an automated system to reduce unnecessary process from receiving up to shipping, creating the additional areas for each warehousing activity, and codifying each of the material in the warehouse to reduce the theft and redundancy of materials in the warehouse.

Keywords: Warehouse, Warehousing management, warehousing performance

Suggested Citation: Suggested Citation

Ararsa Buzu (Contact Author)

Jimma university (ju), college of business and economics ( email ), do you have a job opening that you would like to promote on ssrn, paper statistics, related ejournals, operations management ejournal.

Subscribe to this fee journal for more curated articles on this topic

Other Production & Operations Management eJournal

warehouse operations Recently Published Documents

Total documents.

• Latest Documents
• Most Cited Documents
• Contributed Authors
• Related Sources
• Related Keywords

A literature review of smart warehouse operations management

AbstractE-commerce, new retail, and other changes have highlighted the requirement of high efficiency and accuracy in the logistics service. As an important section in logistics and supply chain management, warehouses need to respond positively to the increasing requirement. The “smart warehouse” system, which is equipped with emerging warehousing technologies, is increasingly attracting the attention of industry and technology giants as an efficient solution for the future of warehouse development. This study provides a holistic view of operations management problems within the context of smart warehouses. We provide a framework to review smart warehouse operations management based on the characteristics of smart warehouses, including the perspectives of information interconnection, equipment automation, process integration, and environmental sustainability. A comprehensive review of relevant literature is then carried out based on the framework with four perspectives. This study could provide future research directions on smart warehouses for academia and industry practitioners.

Towards an increased understanding of learning: a case study of a collaborative relationship between a retailer and a logistics service provider

PurposeThis study aims to explain how learning occurs in collaborative retailer–logistics service provider (LSP) relationships. The research is guided by two research questions, addressing absorptive and desorptive capacities and the interaction between these capacities.Design/methodology/approachThe study is based on a case study of a Swedish, collaborative retailer–LSP dyad. The empirical data are structured around five specific learning situations within the retailer–LSP dyad.FindingsThe findings provide an explanation for how learning occurs within a collaborative retailer–LSP relationship based on subprocesses of absorptive and desorptive capacities. The interaction between these processes is found to rely on two types of support: one-directional and bidirectional. The findings also indicate positive outcomes of learning, such as improved cost efficiencies in warehouse operations, better customer services and improved long-term strategic planning.Practical implicationsThis study shows how retailers and LSPs can learn from each other and together create an improved logistics system for end customers.Originality/valueThis research takes into account absorptive and desorptive capacities in a collaborative retailer–LSP relationship. This study enhances the understanding of inter-organisational learning processes in a retail logistics context.

WAREHOUSE OPERATIONS OF PASSENGER SERVICE CAR DEPOT WITH LOGISTICS ELEMENTS

The article discusses the issues of rational organization of the warehouse process of storing inventory, the sequential implementation of warehouse operations, the improvement of labor organization and technological solutions, the effective use of various equipment involved in the performance of technological operations at the warehouse of the carriage depot of the Joint Stock Company «Uzpasstrans». The issues of creating a database for storing goods, an effective system for organizing their account are considered. The ER-diagram of the database of warehouse inventory of spare parts and components has been developed. The functions and procedures for working with database data on components and spare parts located in the warehouse of the carriage depot have been determined.

A Survey of the Literature on Order-Picking Systems by Combining Planning Problems

Companies have been trying continuously to reduce their logistics costs in the current competitive markets. Warehouses are important components of the logistics systems and they must be managed effectively and efficiently to reduce the production cost as well as maintain customer satisfaction. Order-picking is the core of warehouse operations and an order-picking system (OPS) is essential to meet customer needs and orders. Failure to perform the OPS process properly results in high costs and customer dissatisfaction. This research aims to investigate the state of the art in the adoption of OPS and provide a broad systemic analysis on main operating strategies such as simultaneous consideration of order assignment, batching, sequencing, tardiness, and routing need. This study reviews 92 articles, classifies combinations of tactical and operational OPS problems, and provides guidelines on how warehouse managers can benefit from combining planning problems, in order to design efficient OPS and improve customer service. Combining multiple order-picking planning problems results in substantial efficiency benefits, which are required to face new market developments.

Determination of Performance Indicators for Warehouse Evaluation: A Case of Medium Sized Warehouses in Nakuru Town

Evaluation of warehouse processes is required for decision making purposes and the improvement of warehouse operations. To evaluate warehouse processes, it is essential to identify key indicators in the warehouse operations. This research was intended to identify the significant indicators of warehouse performance that would support management decision making on the improvement of warehouse operations. In this research 20 indicators were identified in four warehouse activities based on Frazelle model. The most important indicators in each warehouse were then determined. The study was conducted through a descriptive design using a survey of 3 medium warehouses in Nakuru. The study employed purposive sampling to select the sample and sample elements. The sample size of 10 warehouse management officers were used in the study. Structured questionnaire was adopted as data collection instrument. The research findings indicated that most important performance indicators for receiving is productivity, for storage is space utilization, for order picking is cycle time and for shipping is productivity. Process improvement steps were proposed based on benchmarking among warehouses handling similar goods. It was noted that different types of warehouses require different indicators. The study concluded that measurement of warehouse processes is important and enables operations improvements and cost reduction. Keywords: Warehouse procedures, warehouse performance, warehouse improvement, medium sized warehouses, Nakuru town, Kenya

Risk Assessment for the Use of Drones in Warehouse Operations in the First Phase of Introducing the Service to the Market

Services, unlike products, are intangible, and their production and consumption take place simultaneously. The latter feature plays a crucial role in mitigating the identified risk. This article presents the new approach to risk assessment, which considers the first phase of introducing the service to the market and the specificity of UAV systems in warehouse operations. The fuzzy logic concept was used in the risk analysis model. The described risk assessment method was developed based on a literature review, historical data of a service company, observations of development team members, and the knowledge and experience of experts’ teams. Thanks to this, the proposed approach considers the current knowledge in studies and practical experiences related to the implementation of drones in warehouse operations. The proposed methodology was verified on the example of the selected service for drones in the magazine inventory. The conducted risk analysis allowed us to identify ten scenarios of adverse events registered in the drone service in warehouse operations. Thanks to the proposed classification of events, priorities were assigned to activities requiring risk mitigation. The proposed method is universal. It can be implemented to analyze logistics services and support the decision-making process in the first service life phase.

Design and application of data analytics in an internet-of-things enabled warehouse

Purpose While several warehouses are now technologically equipped and smart, the implementation of real-time analytics in warehouse operations is scarcely reported in the literature. This study aims to develop a practical system for real-time analytics of process monitoring in an internet-of-things (IoT)-enabled smart warehouse environment. Design/methodology/approach A modified system development research process was used to carry out this research. A prototype system was developed that mimicked a case company’s actual warehouse operations in Indonesia’s manufacturing companies. The proposed system relied heavily on the utilization of IoT technologies, wireless internet connection and web services to keep track of the product movement to provide real-time access to critical warehousing activities, helping make better, faster and more informed decisions. Findings The proposed system in the presented case company increased real-time warehousing processes visibility for stakeholders at different management levels in their most convenient ways by developing visual representation to display crucial information. The numerical or textual data were converted into graphics for ease of understanding for stakeholders, including field operators. The key elements for the feasible implementation of the proposed model in an industrial area were discussed. They are strategic-level components, IoT-enabled warehouse environments, customized middleware settings, real-time processing software and visual dashboard configuration. Research limitations/implications While this study shows a prototype-based implementation of actual warehouse operations in one of Indonesia’s manufacturing companies, the architectural requirements are applicable and extensible by other companies. In this sense, the research offers significant economic advantages by using customized middleware to avoid unnecessary waste brought by the off-the-shelves generic middleware, which is not entirely suitable for system development. Originality/value This research’s finding contributes to filling the gap in the limited body of knowledge of real-time analytics implementation in warehousing operations. This should encourage other researchers to enhance and develop the devised elements to enrich smart warehousing’s theoretical knowledge. Besides, the successful proof-of-concept implementation reported in this research would allow other companies to gain valuable insights and experiences.

An Overview of Warehouse Operations for Cold Chain

An analytical design & optimization approach to enhance warehouse operations, genetic algorithm solution for transfer robot operation.

The proposed article presents the research of decision-making processes for control system of transport robot, acting inside warehouse. The analysis shows the growing importance of warehouse systems for flexible integrated system. Such warehouse operations, like loading/uploading pallets and sending goods to distribution center can be fully automated by applying robots and some examples are known. Simultaneously, the wide introduction of transport robots with extended possibilities is still limited. Application of intelligent robot can significantly improve properties of warehouse systems, making them closer to Industry 4.0 standards. The functioning of a transport robot can be described as a Travelling Salesmen Problem with numerous solutions that must be effectively limited. Genetic algorithms are proposed as a basis for path generation. The algorithms and software are implemented to model warehouse transfer robots using an Arduino-robot. The results of testing show the effectiveness of the developed algorithms, software and hardware system.

Export Citation Format

Share document.

The transformation from manual to smart warehousing: an exploratory study with Swedish retailers

The International Journal of Logistics Management

ISSN : 0957-4093

Article publication date: 22 June 2022

Issue publication date: 19 December 2022

To meet customers' expectations on shorter lead times, high product availability, flexibility, and variation in delivery and return options, retailers have turned their attention to warehousing and are making big investments in technology. Currently, technology providers are pushing for smart warehousing, a new and under-researched phenomenon. This study aims to conceptualize the term and examine pathways toward implementing smart warehousing.

Design/methodology/approach

An exploratory survey was administered to 50 leading Swedish retailers in varying segments. A two-tailed t -test for equality of means was used to detect significant differences between current and future states.

The study found that future smart warehouses will be automated, autonomous, digital, and connected, but that retailers will follow different paths along this journey, driven by contextual trends, e.g. sales growth, wider product assortment, shorter lead-time offerings, and integration of brick-and-mortar and online stores. Interestingly, the study revealed that many of the retailers that aim to create smart warehouses in five years are not the retailers with the most developed technology today.

Research limitations/implications

The paper operationalizes smart warehousing in two dimensions: degree of automation and degree of digitalization and connectivity of information platforms. Based on the findings, 16 theoretical propositions are put forth that, based on contextual factors, explain different pathways for retailers to implement smart warehousing.

Practical implications

The empirical insights and theoretical discussions provide practically useful guidance, including outlined trends, for selecting and benchmarking automation and complementary technologies in warehouse operations.

Originality/value

This paper conceptualizes and operationalizes smart warehousing – an original approach. It is also one of the first to investigate the technological transformation in retail warehousing empirically, explaining how and why retailers choose different pathways toward smart warehousing.

• Transformation

Kembro, J. and Norrman, A. (2022), "The transformation from manual to smart warehousing: an exploratory study with Swedish retailers", The International Journal of Logistics Management , Vol. 33 No. 5, pp. 107-135. https://doi.org/10.1108/IJLM-11-2021-0525

Emerald Publishing Limited

Copyright © 2022, Joakim Kembro and Andreas Norrman

Published by Emerald Publishing Limited. This article is published under the Creative Commons Attribution (CC BY 4.0) licence. Anyone may reproduce, distribute, translate and create derivative works of this article (for both commercial and non-commercial purposes), subject to full attribution to the original publication and authors. The full terms of this licence may be seen at http://creativecommons.org/licences/by/4.0/legalcode

1. Introduction

The retail industry has been undergoing a digital transformation coupled with customers' expectations of shorter lead times (i.e. demanding same-day delivery), high product availability, flexibility when and where to shop, and varying delivery (e.g. click-and-collect, pick-up points, home delivery) and return options ( Galipoglu et al. , 2018 ; Tokar et al. , 2020 ). To meet these demands, the logistics network, particularly the warehouse, has been highlighted as a critical component ( Kembro et al. , 2018 ). The warehouse, previously viewed as a “necessary evil” in the supply chain, now plays a key role in fulfilling customer orders and significantly influences both logistics costs and service levels ( Faber et al. , 2018 ). As Rouwenhorst et al. (2000 , p. 515) put it: “[T]he efficiency and effectiveness in any distribution network … is largely determined by the operation of the nodes in such a network, i.e. the warehouses.”

To improve warehouse operations, online global giants, e.g. Alibaba and Amazon have made large investments in automated material-handling technology ( Alibaba Cloud, 2019 ; Amazon, 2020 ). With increased competition and more mature and varied technologies, the automation trend also has spread among retailers and logistics service providers worldwide ( Reiser, 2020 ; MH&L, 2020 ). An important driver is that the latest generation of automation technologies offers flexibility to handle different types of products and adjust to demand variations, enabling effective and efficient storage, handling, and sorting of large product flows ( Azadeh et al. , 2019 ; Kembro et al. , 2022 ). To improve warehouse operations further, retailers couple automated material handling with digitalization and connectivity of information platforms. Examples of new technologies that are relevant for warehouse operations include artificial intelligence (AI), Internet of things (IoT), cyber-physical system (CPS), big data, 5G, and intelligent video analysis (IVA) ( Kembro et al. , 2017 ; Kamali, 2019 ; Taboada and Shee, 2021 ; Winkelhaus and Grosse, 2020 ; Chung, 2021 ).

This combination of technologies sometimes is termed smart warehousing ( Mahroof, 2019 ; Kamali, 2019 ), a term that is gaining increased attention, often in connection with Industry 4.0, Logistics 4.0, and IoT ( Lee et al. , 2018 ; Winkelhaus and Grosse, 2020 ; van Geest et al. , 2021 ; Issaoiu et al. , 2021 ). It is used in tech blogs, industry reports, conference proceedings, and scientific journal papers. However, extant smart-warehousing research is fragmented when it comes to its substance (see, e.g. Bolu and Korcak, 2019 ; Chung, 2021 ; Zhang et al. , 2021 ). Previous research has instead listed or developed sub-applications of smart technologies that could be used. This gap is important to address to bring research together, facilitate joint discussions, and enable analysis of patterns on a more holistic level, instead of focusing only on specific applications of certain technologies. Another important research gap is the lack of clarity on how to implement smart warehousing ( Azadeh et al. , 2019 ; Winkelhaus and Grosse, 2020 ; van Geest et al. , 2021 ). During the journey toward smart warehousing, companies operate in different contexts and, therefore, are likely to follow different paths ( Kembro and Norrman, 2021 ). However, extant research does not provide any guidance or explanation for different pathways on how and why retailers calibrate their timing, technology, and focus to suit certain operations during their transformation from manual to smart warehousing.

To address these gaps, this study aims to conceptualize smart warehousing and explain pathways on how to implement it. By administering an empirical exploratory survey to 50 Swedish retailers, we make several contributions. Theoretically, we contribute by operationalizing smart warehousing in two dimensions – degree of automation and degree of digitalization and connectivity of information platforms – and by conceptualizing future smart warehouses as automated, autonomous, digital, and connected. We also contribute by identifying different pathways to smart warehousing and explaining how retailers calibrate their timing, technology, and focus to suit certain operations using 16 theoretical propositions. For managers, our study outlines pioneering practices that can help other retailers understand critical issues earlier, as well as how to address them. Our findings provide insights into technologies that are expected to grow in use and criticality to support material handling in single warehouses and in increasingly complex and decentralized networks.

The paper is structured as follows. In Section 2 , we present a theoretical background on warehouse operations in retailing and technology's role in warehouse management. In Section 3 , we describe our empirical study's design. In Section 4 , the findings are presented and analyzed, describing how companies intend to transform from manual to smart warehousing. In Section 5 , based on our findings, we submit theoretical propositions, then conceptualize smart warehousing and explain pathways. Finally, in Section 6 , we outline conclusions and suggestions for future research.

2. Theoretical background

2.1 warehouse operations in retailing.

Warehouses represent “the points in the supply chain where [the] product pauses, however briefly, and is touched” ( Bartholdi and Hackman, 2016 , p. 3). In the retail context, warehouses primarily are used to consolidate and store a range of products to reduce transportation costs and lead times. Different types of retail warehouses include the distribution center (DC), online fulfilment center (OFC), and micro-fulfilment center (MFC). Warehouse operations include receiving, put-away, storage, picking, sorting, packing, and shipping. Along with managing increases in online orders, warehouses also might handle returns and cross-docking, in which goods move directly from receiving to shipping ( Kembro et al. , 2018 ).

In a warehouse, arriving products are checked for quality, registered, and potentially re-packed before being put away in assigned storage locations ( Frazelle, 2016 ). Storage, which includes a reserve and picking area, typically is divided into zones, depending on stock-keeping unit (SKU) characteristics (e.g. size and temperature requirements) and/or order characteristics (e.g. online orders vs store replenishment) ( Eriksson et al. , 2019 ). SKUs can be dedicated or randomly assigned to a location. A common approach is to combine the two (also termed class-based storage ), i.e. SKUs are placed randomly within a dedicated area, thereby reducing travel while avoiding congestion ( Gu et al. , 2007 ). Picking comprises most of the operational cost and has been by far the most-researched warehousing topic. Picking efficiency can be improved by putting the fastest-moving products in the most convenient locations ( Gu et al. , 2007 ) and by selecting appropriate picking methods. The four most common picking methods are single, batch, zone, and wave ( Bartholdi and Hackman, 2016 ). Kembro and Norrman (2019) described the SKU extraction method as one “which implies that a large number of customer orders are accumulated, and the picker only goes to one or a few storage locations and (manually) takes out hundreds of the same SKU to a trolley or pallet. These products are then moved to a sorting system” (p. 521). Eventually, orders are packed and shipped. If an order involves multiple flows (e.g. wave picking or cross-docking), the first step is to sort and merge the various order lines per customer and destination. SKUs are registered thereafter for departure and positioned based on the designated gate and time window ( Bartholdi and Hackman, 2016 ).

To improve warehouse operations' efficacy and efficiency, several configuration aspects are considered, including physical layout (e.g. placement of docks, aisle configuration, and lane depth) and storage and handling equipment (e.g. racks and forklifts for put-away and picking) ( Kembro et al. , 2018 ). Important resources also include labor management (e.g. scheduling, rotation, and shifts) ( De Leeuw and Wiers, 2015 ), information systems (e.g. warehouse management systems [WMS] and warehouse-control systems [WCS]) ( Kembro and Norrman, 2019 ), and automation technologies, e.g. conveyors and robots ( Baker and Halim, 2007 ; Azadeh et al. , 2019 ). These configurations' goal is to improve utilization of resources and capacities (labor, space, and equipment), increase throughput, reduce material-handling time, and increase operations and design flexibility ( Kembro and Norrman, 2019 ). However, which warehouse configuration to use depends on a range of contextual factors ( Hassan et al. , 2015 ; Faber et al. , 2018 ; Eriksson et al. , 2019 ; Kembro and Norrman, 2021 ), e.g. order-fulfillment time requirements, assortment range, number of transactions, and sizes of goods.

2.2 Technology's role in warehouse management

Just a decade ago, warehousing worldwide predominantly were manual operations, but increasing competition and expectations on shorter lead times (demand for same-day delivery), high product availability, and a variation of delivery and return options have necessitated improvements to transform from manual to smart warehousing (e.g. Chung, 2021 ; Kembro and Norrman, 2021 ; Zhang et al. , 2021 ). As described below, this transformation mainly involves technology, e.g. automated material handling and information systems.

2.2.1 Automated material handling

Warehouse automation has been around for many years. Defined as “the direct control of handling equipment producing movement and storage of loads without the need for operators or drivers” ( Rowley, 2000 , p. 38), it includes equipment, e.g. automated storage and retrieval systems (AS/RS), automated guided vehicles (AGVs), and conveyorized sorting systems. Positive aspects of warehouse automation include improved space utilization and service, as well as lower operating costs and reduced picking errors. However, it requires significant investments, and more static automation technologies may involve flexibility risks, e.g. during extreme demand peaks ( Baker and Halim, 2007 ; Kembro and Norrman, 2020 ). Therefore, these investments require careful analysis, considerable scale, and a long-term vision ( De Koster, 2018 ).

In recent years, retailers have increased the pace of implementing automation in warehouse operations, and various new technologies are being tested to make material handling more effective and efficient ( Kembro and Norrman, 2020 ). Azadeh et al. (2019) conducted a comprehensive review of recent developments in various automation technologies in warehousing, identifying the following categories: Crane/Automated Forklift (e.g. AS/RS, push-back rack); Carousels and Dispensers (e.g. horizontal carousel, A-frame); Shuttle (aisle-based, e.g. horizontal AVS/R system; grid-based, e.g. robotic compact storage and retrieval systems); and AGVs (e.g. movable racks in robotic mobile fulfillment systems). One interesting development is the increased flexibility in automation technologies, with different solutions available for different SKU characteristics, while it is also less difficult to adjust to demand variations. It is also relevant to consider whether humans will work in warehouses in the future and what their role will be. Azadeh et al. (2019 , p. 940) noted that: “Human picking in collaboration with AGVs is one of the most recent technologies that is becoming popular in practice because of its simplicity and flexibility.”

2.2.2 Digitalization and connectivity of information platforms

Simultaneously, retailers develop information systems to manage and control their warehouse operations. As Kembro and Norrman (2019) asserted, the three most common systems include enterprise resource planning (ERP) systems; WMS; and WCS. ERP connects information across a range of organizational functions, e.g. sourcing, inventory management, production planning, and financial matters. WMS is used for shorter-term planning and control of resources and order fulfilment for various warehouse operations. WCS is used to control automated systems, e.g. robots and advanced sorting algorithms. A related system is the warehouse execution system (WES), which synchronizes various automation technologies' operation with workers and could be viewed as an integration or combination involving WMS and WCS. Several other systems besides these also exist, e.g. for labelling, administering, and transport administration system (TAS). With the increasing importance of sharing real-time inventory and order information, both internally and externally, it becomes vital for all systems within each warehouse and across the logistics network to be integrated. For this purpose, Kembro and Norrman (2019) noted that retailers increasingly use a distributed order management (DOM) system.

2.3 The transformation toward smart warehousing

Smart warehousing is a term that is gaining increased attention, often in connection with Industry 4.0, Logistics 4.0, and IoT (see literature reviews in Winkelhaus and Grosse, 2020 ; Chung, 2021 ; Issaoui et al. , 2021 ). The term has not been defined precisely, but generally has been described as something more than using automated material handling and traditional information and communication technology (ICT), e.g. ERP, WMS, and WCS. Chung (2021 , p. 1) defined smart technologies as “applications of artificial intelligence and data science technologies, e.g. machine learning, big data, to create cognitive awareness (autonomous) of an object with the support of information and communication technologies, e.g. IoT and blockchain.” Previous research often either has listed different smart technology applications ( Winkelhaus and Grosse, 2020 ; Chung, 2021 ; Issaoui et al. , 2021 ), developed focused algorithms (e.g. Jiang et al. , 2021 ), proposed IoT-based WMS ( Lee et al. , 2018 ; Aamer and Sahara, 2021 ), or referred to architectures for information systems related to data collection and administration ( Jabbar et al. , 2018 ; van Geest et al. , 2021 ).

Several studies have used smart warehousing to describe a warehouse that includes a combination of automated material handling and AI. For example, Bolu and Korcak (2019) discussed smart warehouses in terms of automation technologies that make warehouse robots and systems smarter, including autonomous robots that perform most activities in a warehouse, as well as information systems that keep track of every object's movement in the warehouse. Mahroof (2019) noted that these robots carry out most warehouse work, have Wi-Fi capabilities, and use self-charging batteries and laser-detection technology. Mahroof (pp. 178–179) concluded that “the new automation age is here, whereby industrial robots and computers are being used beyond their traditional scope of performing highly accurate repetitive tasks, routine physical work tasks, through to more complex tasks that require cognitive capabilities, e.g. making tacit judgments, sensing emotion, and driving processes which previously seemed impossible.” Along the same lines, Zhang et al. (2021) investigated Alibaba's smart warehouses, which use a set of AI applications, along with collectively working robots and other related human and organizational resources, to improve key business processes' efficiency and efficacy. The researchers noted that while humans focus on higher-value tasks that require creativity, robots and AI are used to execute repetitive, time-consuming, and/or hazardous tasks. With WMS, WCS, and AGVs as foundational systems, AI uses algorithms as building blocks to automate, augment, or transform processes. Examples of algorithms include sales forecasting, location recommendation, 3D packing, order wave combination, route planning, robot scheduling, and robotic motion control.

Extending this perspective, Kamali (2019) described the smart warehouse as using a mix of technologies, including, e.g. robotics systems, IoT, radio-frequency identification (RFID), enterprise asset management (EAM), and digital twins with 3-D representations of objects and their components, e.g. sensors. Other relevant technologies include wide area network (WAN), cloud computing, automatic identification technology, CPS, and blockchain. Kamali argued that AI allows for gaining unprecedented insight into products, components, and even materials' life cycles, noting that smart warehouses make activities more efficient, save on labor costs, reduce errors, and generate higher productivity. Thus, the automation and AI focus is complemented (e.g. Lee et al. , 2018 ; Kamali, 2019 ; Issaoui et al. , 2021 ) by adding (inter)connectivity in terms of IoT and wide area networks/cloud for information management. Some studies have developed frameworks for WMS in the IoT context ( Lee et al. , 2018 ; Aamer and Sahara, 2021 ) or reference architectures ( Jabbar et al. , 2018 ; van Geest et al. , 2021 ). Another relevant technology is IVA, which enables advanced analytics and decision making in real time in the warehouse ( Kembro et al. , 2017 ). Using a mix of technologies is supported by Attaran (2020) , who pointed out that AI and robotics, cloud computing, 3D printing, advanced analytics, blockchain, AR, RFID, IoT, and cloud technology drive digital trends and elicit change in supply chain management. Similar reasoning can be found across tech blogs and suppliers. For example, Flytware (2021) stated: “Smart warehousing is essentially a set of interconnected and/or automated technologies for streamlining warehousing operations in an efficient manner.” The connectivity dimension links the warehouse as part of a digital network, pointing to the growing importance of Industry 4.0, digitalization, and a unified platform for multiple device connectivity in real time, which can be implemented through 5G networks ( Taboada and Shee, 2021 ). Another aspect sometimes stressed is robot autonomy ( Winkelhaus and Grosse, 2020 ; Chung, 2021 ; Jiang et al. , 2021 ).

To sum up, smart warehousing is receiving increased attention in both practice and academia. but extant research is fragmented, with no consensus reached on a definition of smart warehousing . Previous research instead has listed or developed sub-applications of smart technologies that could be used. This literature gap is important to address to facilitate discourse and create a common understanding of what smart warehousing is. As a starting point, our literature review revealed two smart warehousing dimensions: the development of automated material handling and development of digitalization and connectivity of information platforms. The review also revealed that pathways toward implementing smart warehousing are missing. Previous studies emphasized contextual factors' importance, suggesting that companies may follow different paths. However, extant research has not provided any guidance or explanation as to how and why retailers calibrate their timing, technology, and focus to suit certain operations during their transformation from manual to smart warehouse management.

3. Research design and method

This study aimed to grasp and generate practically relevant insights and theoretical propositions about a new phenomenon – smart warehousing – on which current knowledge is scarce concerning both the current situation and future developments. By illuminating this area for consideration, we seek to influence the definition of its problem domain, leading to contributions of the “theoretical pre-science” type ( Corley and Gioia, 2011 ).

For this, we used an exploratory survey (distinguished from explanatory and descriptive research). It is useful for becoming more familiar with a topic and identifying new possibilities and dimensions of interest ( Malhotra and Grover, 1998 ), as well as identifying interesting patterns during the early stages of the research-maturity cycle ( Malhotra and Grover, 1998 ; Edmondson and McManus, 2007 ), rather than testing a theory-driven hypothesis. The study is theory-elaborative and tries to adapt theory to contemporary practices and challenges ( Corley and Gioia, 2011 ; Ketokivi and Choi, 2014 ). Therefore, conducting an exploratory survey is appropriate for revealing trends, interesting patterns, and facets of the phenomenon with the purpose of developing propositions for future research studies. In this way, exploratory surveys (conducted with experts in the field) are similar to case studies and other qualitative methods, which also are useful for discovering new facets of phenomena under study ( Forza, 2002 ). Exploratory surveys have been used in supply chain and operations management research before, e.g. related to information systems: Themistocleous et al. (2001) examined problems related to ERP systems, and Kembro and Norrman (2019) investigated omnichannel logistics issues. These studies applied more open-ended research questions in combination with closed-ended survey questions using ranking or Likert scales.

3.1 Data collection

This study is the first part of a larger study, the Swedish Retail Logistics panel, in which Swedish retail companies (pure-play online retailers, henceforth referred to as e-tailers , as well as retailers with both online and brick-and-mortar stores) have been invited to participate in a series of exploratory surveys. The retailers were identified through addresses from different databases and listings of leading Swedish retailers (e.g. related to growth or turnover), e.g. the report “Who is who in Swedish retail 2020” ( Lindecrantz, 2020 ). Altogether, 50 retail companies provided input for this study (out of 300-plus companies who were sent invitations). The retailers represent a wide range of retail sectors and product types (see Table 1 ; respondents could provide multiple answers). The major retail sectors, currently at the forefront of online sales, are well-represented by their leading companies in the survey. Most of the retailers that are part of the panel are top-five (Sweden) in their segments based on turnover.

Our empirical data collection is based on an online survey. The respondents were senior supply chain/logistics managers divided between the following positions: 37 directors (or heads) of supply chain management/logistics/operations; one CEO; four warehouse managers; five managers responsible for development/design of logistics and/or warehouse solutions; one innovation lead; one head of e-business; and one head of an online store. The survey was pre-tested on five company representatives to test the questions' general appropriateness, functionality, and structure. The feedback was used to modify the survey instrument, mainly by increasing clarity in survey questions and explaining terminology. The respondents were asked to use a perceptual Likert scale ranging from 1 (“agree to a very low degree”) to 7 (“agree to a very high degree”) to assess their current focus and development in different areas. To address the limitations of scales in an exploratory study, most questions that included statements with fixed-scale alternatives were complemented with open-ended questions to explore alternative answers on the specific topic.

The main survey questions focused on the following issues: (1) company data and contextual factors, e.g. size, turnover, type of products, type of channel, assortment range, and lead-time offerings; (2) to what degree the companies invest in and use automation in warehouse operations – both in general and for different material-handling processes; (3) to what degree companies use technologies (other than automation) in the warehouse, e.g. regarding information platforms and connectivity. The respondents were asked to list and explain their three main focus areas related to the implementation of warehouse automation. Some questions included pre-determined answering alternatives, e.g. different types of technologies (with clarifying examples to reduce the risk of misunderstanding). We used many different sources to develop the survey questions, including scientific literature (discussing warehouse operations in retailing, warehouse automation and technology, and smart warehousing and technologies) and gray literature, e.g. business journals and tech blogs. We also scanned the market to understand more about automation technology and smart warehousing that either already exist or are being developed. As we initially looked for responses from both very small e-tailers (with large growth) and larger traditional and omnichannel (i.e. integrated store and online) retailers, we defined scales using a logarithmic approach to illustrate the large difference between potential answers regarding company data and contingency factors.

3.2 Data analysis

With an exploratory aim and too few respondents to conduct advanced statistical analysis, we described patterns in the participants' perceptions of current and future practice. We were inspired by multiple case study analyses (see, e.g. Miles and Huberman, 1994 ), with the aim of developing propositions for future research by pattern-matching data from open-ended and closed-ended questions. The same set of questions was used for both current and future states, and to detect significant differences in any context, we used a two-tailed t -test for equality of means.

To support the formulation of theoretical propositions about different pathways toward smart warehousing, we analyzed the data based on different contextual factors. First, we clustered the data based on channel strategy (with today's share of retailers stated): e-tailing (may include showrooms) (26%); partly integrated multichannel (both physical stores and online sales, but separated or only partly integrated channels) (50%); omnichannel (high degree of integration of store and online channels) (16%); and store focus, without (or locally store-driven) online sales (8%). Interestingly, our data indicate a strong trend (in five years) toward implementation of an omnichannel strategy among the panel retailers. Retailers selling only offline decreased (which is not a surprise), while store-driven online sales increased. Another observation is that e-tailers continued to focus on online sales and, with one exception, did not plan to open physical stores. The other contextual factors that we focused on in our analysis included turnover, assortment range, and sizes of goods.

4. Findings and analysis

To understand how Swedish retailers intend to transform from manual to smart warehouse management, we examined current trends and future intentions on several aspects. We present these findings in the following sub-sections and use them to develop 16 theoretical propositions.

4.1 Overall degree of automated warehouse operations

The study indicated an increased willingness to invest in automation, with retailers making large investments (>SEK 100 million annually), more than doubling such expenditures (8–22%), while those investing SEK 31–100 million annually increased such expenditures from 12 to 22%. Retailers not investing at all in automation decreased from 32 to 8%, indicating a statistically significant (see t -tests in Appendix ) and a strong increase in the overall degree of automation (in the warehouse with the most recent investment). On a scale from 1 (very low degree) to 7 (very high degree), the mean values increased from 2.07 (five years ago) via 3.15 (at present) to 5.13 (in five years). The retailers represented four clusters. The first group of retailers made large investments in recent years, but currently focuses on fine-tuning automation technology (e.g. by supplementing it with other technology). The second cluster has not yet automated to a large extent but intend to invest significantly in automated systems in the coming years. A third group represents retailers who already have made significant investments and continue to invest heavily in warehouse automation in their logistics networks. The fourth cluster, which is decreasing in number, represents retailers that have not and do not intend to automate their warehouses.

The degree of automation was analyzed on a deeper level, related to different warehouse operations ( Figure 1 ). The operations that have been automated to a greater extent represented outbound flows. Above all, “sorting outgoing goods” (average 3.30) stood out, while picking (2.89) and packing (2.80) also were higher, indicating that the most labor-intensive processes were automated first to justify return on investment. The focus on sorting also reflected the increased extent, variety, and complexity of sorting at the operational material-handling node level ( Kembro et al. , 2022 ). However, on the scale from 1 to 7, values below 4 still indicated a low degree of automation. The operations that currently are automated to a very low degree (thereby still requiring a greater share of manual work) are goods receipt (1.61), sorting of incoming goods (1.63), and handling returns (1.50). These flows included a wide variety of goods and requirements for quality control before storage. It is also important to determine the storage zone and balance incoming flows to avoid bottlenecks. Previous research (e.g. Kembro and Norrman, 2019 ) has indicated that retailers often position their most experienced staff in goods receipt.

Five years from now, a significant increase (see t -tests in Appendix ) in automation is expected, with several operations expected to average above 4.00 (storage 4.24, picking 4.98, packing 4.89, and outgoing sorting 5.32), indicating a clear trend toward a relatively high degree of automation (i.e. not only an increase, but also an increase to a relatively high level). Despite current low levels, the trend toward a higher degree of automation for incoming goods is interesting. Insights from a previous study ( Kembro and Norrman, 2019 ) indicate that automation of inbound processes requires greater work and coordination with suppliers, e.g. standardizing boxes/labels and balancing flows handled in warehouses (to avoid bottlenecks). It also seems that retailers want to try to make return handling more efficient, with some retailers stating that “automation of return flows is one of our top-three focus areas for technology in warehousing.”

Analyses of individual retailers and the number of warehouse operations that one has stated as highly automated (response alternatives 6–7 on a scale of 1–7) in five years provide additional insights ( Table 2 ). Twenty percent of the retailers answered that they will have automated five or more operations to a very high degree. Of these, many will have automated most of the warehouse operations to a high degree, and some even plan to use a fully automated warehouse. Altogether, 44% indicated that they will automate one to four operations significantly. Simultaneously, many retailers (38%) have not specified any of their operations as being highly automated, indicating that a high demand for manual work will remain in warehouses, but this differs between different retailers.

The more labor-intensive the processes, the higher the priority for retailers to automate. This implies that retailers begin by automating outbound picking, packing, and sorting (and related storage), followed by shipping and cross-docking to focus ultimately on inbound receipt, sorting, and handling returns.

4.2 Contextual factors' influence

To understand differences between the retailers, we first analyzed the data based on channel strategy. This analysis revealed that today's omnichannel retailers automated early (most indicated values 4–7 already five years ago), while e-tailers and retailers with partially integrated multichannels caught up by increasing the degree of automation sharply for the past five years. Retailers with a store focus remain at a low level. On a more detailed level, future omnichannel retailers, to some extent, will lead automation of incoming goods (receipt of goods, sorting of incoming goods, storage). E-tailers, to some extent, have less “cross-docking” compared with other retailers (which is logical considering that incoming pallets and cartons must be broken down for picking for e-customers). Instead, to a greater extent, they have invested in “sorting outgoing goods.” The greatest uncertainty (answer alternative “do not know”) applies to cross-docking and return handling, which, in itself, is an interesting observation.

Second, we analyzed the retailers' turnover. Most of the retailers in the panel (94%) currently have a turnover larger than 100 MSEK (∼10 MEuro), of which 36 (72%) sell more than one billion SEK (100 MEuro). Two of the retailers are global giants and sell over 100 billion SEK (∼10 billion Euro). Thus, relatively large retailers (rather than medium-size and small companies) dominate the panel. In five years, both the smaller and larger retailers plan quick growth. The analysis indicates that retailers with low turnover tend not to automate, which is in line with the large capital/investments required for automation. We noted a correlation between high turnover and high degree of automation ( Figure 2 ). An interesting observation is that medium-size companies presently have a relatively low degree of automation, but the tendency is that they are taking a big leap in five years.

Third, we analyzed assortment range, which generally continues to increase. Partially integrated multichannel retailers have driven this development (of which most aim to be highly integrated omnichannel retailers within five years). Altogether, 68% of these retailers will have an assortment range larger than 30,000 articles. The analysis indicates that retailers with a relatively small assortment tend to maintain a low degree of automation throughout their warehouses, while those with a larger assortment range (which are dominant in the study) increase their degree of automation the most ( Figure 3 ).

Fourth, we analyzed sizes of goods ( Figure 4 ). Presently, this mainly entails handling very small (0–1 liter) and small (1.1–50 liter) goods with automation. In five years, the trend is toward a significant increase in all sizes of goods (with the smallest for bulky, i.e. goods that do not fit on EUR pallets), with high averages for very small/small goods (5.60/5.39). Automated handling of medium-size goods (51–200 liters) also has been increasing (from 1.94 to 3.38). For large goods (>200 liters, which can fit on EUR pallets), a significant increase was found, but still at a generally low level. Among the answers were several “Do not knows,” which, in this case, also represent “not applicable,” a question that is not relevant to the retailer if it does not handle goods of this size in its warehouses.

Omnichannel retailers increased automated handling of large and complex order and goods flows early, but as other retailers catch up on their investments, channel strategy will have less explanatory value on the overall degree of automation.

The larger the turnover and assortment range, and the smaller the sizes of the goods, the more retailers invest in automation technology to improve material handling throughout the warehouse.

4.3 Choice of automation technology for material handling

Another interesting aspect is the choice of automation technology for material handling ( Figure 5 ). We observed generally rare use (Answer Option 1) of most examined automation technologies, but some stood out and are significantly more common than the others: stationary automated sorting systems (3.50); stationary automated storage and retrieval (AS/RS) (2.93); compact/grid-based storage (3.12); and automated packaging systems (2.88). Interestingly, for these technologies, the situation is dipolar , i.e. retailers did not use them at all or rarely. This is explained through our analysis: Some automation technologies tend to be chosen for a certain context (e.g. product characteristics), while others are used more generally among retailers. For storage and picking, companies mainly choose an automation technology depending on goods/flow characteristics: Omnichannel retailers with a mix of goods and order characteristics (e.g. both store replenishment and online) typically choose stationary AS/RS storage, while e-tailers typically choose compact/grid-based storage (e.g. AutoStore) to handle large assortments of relatively small goods (e.g. clothes) that individual online customers order. Retailers that handle piece-pick-intense online orders (i.e. large volumes of small goods) opt for A-frame automatic dispensers. Simultaneously, there tends to be a more general approach in terms of automation of packaging, weighing, dimensioning, sorting, and palletizing of outgoing goods (though mainly retailers with store networks use automatic palletizing).

In five years, use of most of these automation technologies will increase, a pattern in line with the current situation observed. We detected significant increases ( t -test; see Appendix ) in stationary automated sorting (5.06), automated packaging (5.06), compact/grid-based storage (4.56), cubing (4.91), robotic palletizing (4.41), stationary automated storage and retrieval (3.72), A-frame systems (2.58), robotized piece-picking (2.25), automated guided vehicles (2.34), and self-driving flexible forklifts (2.00). Several technologies reached over 4.00 (average), and most are dipolar , i.e. while some retailers are strong advocates (blue), others will not implement such technology at all (black/red). This implies that retailers will opt for different technologies as they automate material handling in their warehouses. However, the future is not as dipolar as the current situation (the color scale is more gradual). One possible explanation is that when retailers' warehouses (e.g. turnover and assortment ranges) grow, they must handle a larger mix of orders, flows, and goods, with more varied characteristics. Thus, retailers introduce multiple zones and may benefit from using a mix of different technologies (e.g. one technology designed to handle smaller goods in one zone and another technology tailored for larger goods in another zone).

The less a warehouse activity is influenced by contextual factors – which mainly applies to inbound and outbound due to more standardized processes and handling units – the more standardized automation technologies are used to improve material handling.

The more a warehouse's activity is influenced by contextual factors – which mainly applies to storage and picking adjusted to, for example, SKU and order characteristics – the more tailored the automation technologies used to improve material handling.

The larger the assortment of relatively small goods, the more retailers automate by using compact/grid-based storage and goods-to-person technology to improve space utilization and increase efficiency of storage and retrieval activities.

The higher the volume of small goods for individual customers, the more retailers automate by using A-frame technology to reduce costs and lead time for piece-picking activities.

The higher the volume of mixed goods and integrated channels, the more retailers automate by using stationary automated storage and retrieval technology to increase space utilization, and the greater the efficacy and efficiency in handling large throughputs of more varied order and product flows.

An interesting observation is that many retailers choose static, rather than flexible, automation technologies mainly for storage, picking, and sorting. This may mean less leeway to change and adapt automation technologies in the future – particularly considering that retailers are making very large investments now and that the future budget for major changes may be limited. Thus, this trend could mean that automation/warehousing, to a greater extent, will dictate conditions in the future (e.g. setting boundaries or creating opportunities) for changes in product assortment, logistics networks, and overall logistics strategy. If retailers invest in scalable and flexible solutions, it may be easier to grow with acquisitions, add more varied store formats, or create more flexible offerings and a wider variety of products. From a historical perspective, large investments in (static) automation technologies were made in the early 2000s, influencing later strategic decisions related to, for example, the designing of networks and warehouses to handle increasing numbers of online orders.

Investments in static automation technology for storage, picking, and sorting reduce a retailer's ability to adapt warehouse operations to future contextual changes, e.g. reduced customer order lead times and wider variety of product offerings.

Today's large investments in static automation technology for storage, picking, and sorting imply that warehouse operations to a greater extent may dictate conditions for changes in product assortment, logistics networks, and overall logistics strategy decisions.

4.4 Complementary technologies in warehouses

The survey also examined other technologies that complement automated material handling ( Figure 6 ), e.g. those related to digitalization and connectivity. Today, WMS is used frequently (average 5.98) and is viewed, more or less, as standard for controlling and managing daily operations in warehouses. Corresponding with WMS, WCS (3.71) and WES (3.07) are used to control and coordinate a variety of processes and automation. In addition to these, pick-by-voice (3.13) and pick-by-light/put-to-light (2.37) are used to some extent.

Digitalization and connectivity technologies, e.g. private networks (4G, 5G) and AI, are used to a small extent (2.74 and 2.25, respectively). However, some individual retailers have invested to a large extent and may be viewed as pioneers. Relatively speaking, more e-tailers than omnichannel retailers submitted high values (6, 7) for WMS, WCS, WCE, private networks, and AI. Today, other technologies are not used at all in principle, including RFID technology (1.08), pick-by-vision (1.29), and IVA (1.12), nor are other hyped technologies used, e.g. the IoT, digital twins, drone technology, or blockchain.

The trend in five years indicates that WMS (average 6.62), WCS (5.95), and WES (5.34) will become the backbone of warehouse management, with large significant increases in WCS and WES. Strong future trends also indicate significant increases for several technologies that humans use to strengthen workers' abilities, e.g. to pick faster or reduce picking errors (so-called “human augmentation”), including put-to-light/pick-by-light (4.45), pick-by-voice (3.16), and pick-by-vision (2.05). The increasing use of these technologies indicates that there still will be human workers in the future who conduct material-handling activities (particularly picking) in warehouses. The sharp rise in use of put-to-light/pick-by-light partly points to the importance of faster labor with lower error rates.

WMS, WCS, and WES represent the backbone of warehouse management, whereas hyped technologies – e.g. RFID, IoT, drone technology, and blockchain – will play a limited role in warehousing in the years to come.

The bigger the focus on online customers, the more retailers invest in technologies that support data management, connectivity, and real-time analysis.

5. The pathway toward smart warehousing

Building on our findings and analysis, we sought to understand to what degree smart warehousing was a tendency or trend among the panel's retailers. With support from the literature, we used survey data to operationalize two dimensions: (1) degree of automation and (2) degree of digitalization and connectivity of information platforms. Degree of automation focuses on automation of material handling, i.e. the handling of physical goods. Examples include stationary automated storage, compact/grid-based storage, A-frame systems, automated sorting systems, and automated weighing and dimensioning. Degree of digitalization and connectivity of information platforms focuses on technologies for handling, analysis, and coordination of information and includes, for example, WMS, WCS, WES, AI, IoT, and private networks.

For each dimension, we summarized the number of technologies (per dimension) for each retailer, in which a high degree of implementation was indicated (Values 5, 6, and 7). We then plotted this value for each retailer in Figure 7 , indicating each company's current position (yellow square) and its intentions in five years (red circle). Automation of material handling is illustrated on the Y -axis, with degree of digitalization and connectivity of information platforms on the X -axis. As previously described, several retailers increased automation of material handling (moving upward along the Y -axis). Simultaneously, many retailers are investing in information platforms and increased digitalization (moving to the right). As the gray arrows illustrate in Figure 7 , the overall trend is that the technology frontier moves diagonally upward to the right, implying that retailers to varying degrees are investing in both automation technologies and information platforms, representing a general tendency toward an intentional technological shift in retailers' warehouses over the next five years.

We conducted additional analyses of the individual retailers' movements. Interestingly, the retailers that perceive themselves as having smart warehouses in five years are not the retailers that are most automated or digitalized today. Instead, several retailers are planning major technological upgrades over the next five years, i.e. moving from having limited technology in warehousing to being at the forefront of development toward smart warehousing. We illustrated this technology shift with four retailers' movements (the blue dotted arrows in Figure 7 ).

Our analysis further indicates that retailers follow different paths toward smart warehousing. On one hand, some retailers follow an automation-focused path toward smart warehousing, i.e. while investing in a wide range of technologies to implement smart warehousing, these retailers emphasize technologies that automate material handling of their physical goods flows. This may arise from the need to manage more varied warehouse operations ( Kembro and Norrman, 2020 ). Examples include handling online customers vs store replenishment; a variation in flows, including handling returns and cross-docking; and large variations in SKU sizes (e.g. pieces, cartons, and pallets). The more varied the operations, the wider the range of automation technology needed for material handling. Another driver is increased sorting complexity (due to, e.g. multiple destinations, delivery modes, and transporters), creating a need for additional automation technology ( Kembro et al. , 2022 ). The historical footprint is also relevant, in which retailers that automated certain material flows early (e.g. store replenishment) add automation technologies dedicated to meeting continuously changing online customers' requirements ( Eriksson et al. , 2022 ).

On the other hand, our data indicate that some retailers, particularly e-tailers, will follow a more digitalization-focused path toward smart warehousing. While these retailers invest in automated material handling, they emphasize information platforms and technology that enable real-time data analysis (e.g. AI). These retailers have less variation in operations and, therefore, require a narrower range of automation technologies for material handling. For example, an e-tailer may have a high automation degree, but can handle warehouse operations with only one main form of automation technology (e.g. AutoStore). This enables a greater focus on other complementary technologies for digitalization and connectivity of information platforms. An important driver is online sales, in which e-tailers generally are more focused on virtual contact with customers, requiring a range of integrated IT systems. Another driver is the need to connect multiple material-handling nodes (e.g. DC, OFC, retail stores) in the logistics network ( Kembro and Norrman, 2019 ).

The advancement of the smart warehousing frontline is driven not by retailers with a current high degree of automated material handling, but rather by pioneering retailers that make a major technological shift from limited use of technology.

Retailers with different channel strategies take different implementation routes toward smart warehousing, in which more-integrated omnichannel retailers follow an automation-focused path, whereas less-integrated retailers and e-tailers follow a more digitalization-focused path.

5.1 Conceptualizing smart warehousing

At this stage, it is also relevant to define smart warehouse (which is currently missing in the literature). Based on extant literature ( Section 2 ) and on an analysis of retailers' current and intended technological implementation, we conceptualize future smart warehouses as: Automated , i.e. robots will handle a large part of physical material handling; Autonomous , in which robots make their own decisions regarding task distribution (e.g. order management) and movements in the warehouse – a combination of autonomous automation also can be called autonomization; Digital , i.e. integrated information platforms handle warehouse management (e.g. inventory levels, sequencing of order picks), including functionality for analysis of large amounts of data (AI/machine learning), e.g. for improved forecasting; and Connected , in which moving resources and products are monitored, controlled, and coordinated in real time. It also enables real-time analysis (e.g. via IVA) of in-store activity to allow for fast decision-making and further development of processes.

We summarize these insights in Figure 8 , which outlines both the two dimensions of smart warehousing, as well as the different stages and pathways toward this goal.

5.2 Connecting multiple smart warehouses in extended logistics networks

Moving beyond the single smart warehouse, retailers will use multiple material-handling nodes in their future logistics networks ( Hübner et al. , 2022 ; Kembro et al. , 2022 ). Our findings indicate that retailers will expand from zero or one to between two and five large distribution warehouses. Our study also found that retailers are adding more and varied material-handling nodes (e.g. DC, OFC, MFC) in their decentralized logistics networks. These are complemented by transformed physical stores, which are becoming the center of retail operations ( Hübner et al. , 2022 ).

Apart from market expansion, the main reason for expanding the number of material-handling nodes is the extremely short lead times from customer order to final delivery. Global giants, e.g. Alibaba and Amazon largely have driven this development, with their increasingly competitive promises to customers ( Kembro et al. , 2022 ). In our study, 52% of multichannel retailers (of which most aim to be highly integrated omnichannel retailers within five years) intend to offer standard lead times under 24 h. A similar pattern is visible for e-tailers. No matter how fast a central warehouse fulfills an order, the transportation times to final destination (e.g. home delivery, C&C) may result in lead times that exceed customer expectations. This development will require well-coordinated logistics networks (e.g. use of drop-shipment and small-scale warehouses, e.g. OFCs or MFCs in and around cities, i.e. closer to end customers). It also will drive the need for effective and efficient material handling across network nodes, which can be implemented, e.g. by investing in new automation technology, as well as advanced and integrated information systems. As Kembro and Norrman discussed ( 2019 ), future warehouses and stores, to a greater extent, will be interconnected, among other ways, through the use of a so-called DOM system. For example, inventory levels are coordinated between different material-handling nodes, and an online order can be routed to/managed in different nodes depending on several defined parameters/goals (e.g. reducing lead times or lowering handling costs).

With increasingly competitive lead-time promises to customers, retailers use more and varied (smaller, localized) material-handling nodes that need to be interconnected in smart warehouse networks.

With requirements on effective and efficient warehousing across logistics networks, retailers increasingly use automated material-handling technology in different types of decentralized material-handling nodes, e.g. micro-fulfillment centers.

6. Conclusions and future research

This study aimed to conceptualize the term smart warehousing and explain pathways on how to implement it. By empirically studying this novel phenomenon, our research influences the definition of its problem domain and offers multiple contributions of the “theoretical pre-science” type ( Corley and Gioia, 2011 ).

Contributing to recent and limited literature on smart warehousing ( Azadeh et al. , 2019 ; Mahroof, 2019 ; Chung, 2021 ; Zhang et al. , 2021 ), we put forth 16 propositions related to automation and complementary technology, as well as pathways toward smart warehousing. Our analysis indicates that the future smart warehouse will be automated, autonomous, digital, and connected, but that retailers will follow different paths along this journey. To support our analysis, we operationalized smart warehousing into two dimensions: degree of automation and degree of digitalization and connectivity of information platforms. This is an important contribution to the literature in different fields (e.g. logistics, operations research, and information systems) that mention smart warehousing without defining it. Our operationalization also could influence future conversion of smart warehousing, enabling analysis of patterns on a more holistic level and focusing not just on specific applications of certain technologies that characterize much of current research.

Interestingly, our study revealed that many of the retailers that aim to create smart warehouses in five years are not the retailers with the most developed technology today. In this transition, retailers followed different technological pathways driven by contextual trends, e.g. the growth of sales, wider product assortment, shorter lead-time offerings, and channel strategy. By demonstrating how the continuously evolving retail landscape influences back-end logistics, we contribute to the literature on retail logistics and warehousing (e.g. Galipoglu et al. , 2018 ; Kembro et al. , 2018 ), as well as related automation technology and information platforms ( Kembro and Norrman, 2019 ). Specifically, we explain why retailers calibrate their timing, technology, and focus to suit certain operations. The study found that retailers first automate labor-intensive, outbound warehousing operations, with an emphasis on small or very small goods. Next, many automate larger-size goods and expand their focus to include inbound operations. An important observation is that automation of outgoing flows is more non-contextual (i.e. similar across retailers), while storage and picking technologies seem more tailored to contextual factors (e.g. characteristics of goods). We also conclude that although new automation technology is available for a wider range of retail segments and sizes, it still requires a large investment. This may explain why retailers with larger turnovers and assortment ranges invest more in automation.

Our study proposes this and provides explanations as to why some retailers focus on advanced automation technology while others tend to pioneer digitalization and connectivity. Retailers generally have a solid understanding of information systems for automated material handling, but have limited knowledge about emerging smart warehousing technologies related to digitalization and connectivity. In five years, WMS, WCS, and WES will be the backbone of warehouse operations, complemented with technologies that support data management and real-time analysis, including AI, IoT, RFID, and IVA. This technological development also is important for connecting logistics networks with multiple, different, and decentralized material-handling nodes (e.g. automated MFCs) to meet growing demand for very short lead times from placed order to delivery.

This study provides practically useful guidance for managers by outlining what is trending now and five years down the road. In many companies and countries, the transformation toward smart warehousing has only just begun. Empirical insights from pioneering practice can help other retailers understand critical issues earlier, as well as how to address them. Our findings provide insights into technologies expected to grow in use and criticality to support both material handling in single warehouses and increasingly complex and decentralized networks. Managers also can use our 16 propositions to reflect on what the near future holds and use them as input for scenario analysis.

Pre-science studies' observations naturally elicit speculation that needs more research. We noted that, related to automation, retailers' current tendency to invest in static automation solutions could limit their future strategic options. Future research could investigate whether these kinds of technological investments follow existing strategy – or whether they instead are driving or delimiting future strategies (e.g. to be able to grow, we need to automate vs our current automation technology, which restricts/supports our scaling up). Explanations as to why some retailers seem to lead the digitalization and connectivity journey should be studied: Are fewer capital investments (compared with automation) required? Do they have fewer nodes and simpler flows on which to focus? Have they reached a higher maturity level regarding information technology?

The conceptualization and operationalization of smart warehousing can be developed further through additional empirical evidence collected in other markets. To build theory, in-depth case study research could be employed to better understand different contingency factors' influence. Of special interest would be research on implementation and transformation (using theoretical lenses, e.g. dynamic capabilities or technology adoption models), economic assessment of investment and performance, and an examination of barriers and opportunities related to the interaction between human and smart technologies in future warehouses. The literature presented mixed perspectives on humans' role in future warehousing. Some warehouse operations remain difficult to automate and may need to be carried out manually ( Azadeh et al. , 2019 ). While some researchers trumpet their unmanned warehouses as a defining characteristic or goal for smart warehouses ( Aamer and Sahara, 2021 ; Jiang et al. , 2021 ), others see robots and AI eventually replacing humans ( Jabber et al. , 2018 ). Some have argued that humans will not be replaced, but rather supported ( Winkelhaus and Grosse, 2020 ), with the intent to better connect people, objects, and physical systems ( Lee et al. , 2018 ). Thus, future research could study which factors explain to what extent future warehouses will be manual, automated, or smart.

Like most research designs, this study has limitations. The sample of retailers (50) that answered the survey was relatively small, not random, and only included retailers from one country. However, we argue that the sample is sufficient for developing propositions regarding the researched phenomenon ( Forza, 2002 ) because Sweden is among the leaders in online sales, and among the top-10 industrial digital transformation countries in 2020 ( Top 10 industrial digital transformation countries in 2020 | InfotechLead ). Furthermore, the most important product segments are covered, including leading retailers within each segment. To pinpoint theoretical and managerial implications further, our research needs to be complemented by and tested through more research. Specifically, the 16 propositions can be tested as hypotheses in future research with more empirical evidence by expanding testing to other markets, both in larger countries at similar stages of transformation toward omnichannels (e.g. the US, UK, and Germany) and in countries that are developed in terms of online sales. To understand smart warehousing technologies, logistics service providers and industrial companies also should be examined. Due to their deeper backgrounds with Industry 4.0 and their connection to smart production, industrial companies might make investments and implement smart warehousing differently than retailers.

In conclusion, the pace of development toward smart warehousing will increase in the coming years. Various systems and technologies will be developed and integrated within and across various material-handling nodes, providing many opportunities for researchers to examine and analyze new challenges and solutions, creating new knowledge in warehousing and retail logistics. Only the future can tell us how smart warehouses evolve and why.

Degree of automation per process, today and in five years ( n  = 50)

Overall degree of warehouse automation over time, related to turnover

Overall degree of warehouse automation over time, related to assortment range

Automation of different sizes of goods, today and in five years

Choice of automation technology for material handling, today and in five years

Other complementary technology used in warehouses, today and in five years

Technology shifts in retailers' warehouses over the next five years; several retailers may hide behind the same point in the chart

Conceptualization of smart warehousing

Multiple, connected smart warehouses in expanding, decentralized logistics networks

Retail sectors and product categories in the panel

Number of highly automated processes per company five years ahead

T-tests to detect significant differences between current and future states

Funding : The Swedish Retail and Wholesale Council funded this research study.

Aamer , A.M. and Sahara , C.R. ( 2021 ), “ Real-time data integration of an internet-of-things-based smart warehouse: a case study ”, International Journal of Pervasive Computing and Communications , Vol.  ahead-of-print No. ahead-of-print , doi: 10.1108/IJPCC-08-2020-0113 .

Alibaba Cloud ( 2019 ), “ Double 11 logistics: Cainiao's battle of ‘billions of parcels’ ”, available at: https://www.alibabacloud.com/blog/double-11-logistics-cainiaos-battle-of-billions-of-parcels_595648 ( accessed 17 September 2020 ).

Amazon ( 2020 ), “ Fulfilment in our buildings ”, available at: https://www.aboutamazon.co.uk/amazon-fulfilment/fulfilment-in-our-buildings ( accessed 15 September 2020 ).

Attaran , M. ( 2020 ), “ Digital technology enablers and their implications for supply chain management ”, Supply Chain Forum: An International Journal , Vol.  21 No.  3 , pp.  158 - 172 .

Azadeh , K. , De Koster , R. and Roy , D. ( 2019 ), “ Robotized and automated warehouse systems: review and recent developments ”, Transportation Science , Vol.  53 No.  4 , pp.  917 - 945 .

Baker , P. and Halim , Z. ( 2007 ), “ An exploration of warehouse automation implementations: cost, service, and flexibility issues ”, Supply Chain Management: An International Journal , Vol.  12 No.  2 , pp.  129 - 138 .

Bartholdi , J.J. III and Hackman , S.T. ( 2016 ), Warehouse and Distribution Science , Georgia Institute of Technology , Atlanta, Georgia .

Bolu , A. and Korcak , Ö. ( 2019 ), “ Path planning for multiple mobile robots in smart warehouse ”, 2019 IEEE 7th International Conference on Control, Mechatronics, and Automation , Delft .

Chung , S.-H. ( 2021 ), “ Applications of smart technologies in logistics and transport: a review ”, Transportation Research Part E: Logistics and Transportation Review , Vol.  153 , 102455 , pp.  1 - 17 .

Corley , K.G. and Gioia , D.A. ( 2011 ), “ Building theory about theory building: what constitutes a theoretical contribution? ”, Academy of Management Review , Vol.  36 No.  1 , pp.  12 - 32 .

De Koster , R. ( 2018 ), “ Automated and robotic warehouses: developments and research opportunities ”, Logistics and Transport , Vol.  38 , pp.  33 - 40 .

De Leeuw , S. and Wiers , V.C.S. ( 2015 ), “ Warehouse manpower planning strategies in times of financial crisis: evidence from logistics service providers and retailers in the Netherlands ”, Production Planning and Control , Vol.  26 , pp.  328 - 337 .

Edmondson , A.C. and McManus , S.E. ( 2007 ), “ Methodological fit in management field research ”, Academy of Management Review , Vol.  32 No.  4 , pp.  1155 - 1179 .

Eriksson , E. , Norrman , A. and Kembro , J. ( 2019 ), “ Contextual adaptation of omnichannel grocery retailers' online fulfilment centers ”, International Journal of Retail and Distribution Management , Vol.  47 No.  12 , pp.  1232 - 1250 .

Eriksson , E. , Norrman , A. and Kembro , J. ( 2022 ), “ Understanding the transformation toward omnichannel logistics in grocery retail: a dynamic capabilities perspective ”, International Journal of Retail and Distribution Management , Vol.  ahead-of-print No. ahead-of-print , doi: 10.1108/IJRDM-10-2021-0508 .

Faber , N. , De Koster , R.B. and Smidts , A. ( 2018 ), “ Survival of the fittest: the impact of fit between warehouse management structure and warehouse context on warehouse performance ”, International Journal of Production Research , Vol.  56 Nos 1-2 , pp.  120 - 139 .

Flytware ( 2021 ), “ Smart warehouse: the present and the future of warehousing operations ”, available at: https://flytware.com/blog/smart-warehouse ( accessed 25 May 2021 ).

Forza , C. ( 2002 ), “ Survey research in operations management: a process-based perspective ”, International Journal of Operations and Production Management , Vol.  22 No.  2 , pp.  152 - 194 .

Frazelle ( 2016 ), World-Class Warehousing and Material Handling , 2nd ed. , McGraw-Hill Educational , New York .

Galipoglu , E. , Kotzab , H. , Teller , C. , Yumurtaci Hüseyinoglu , I.Ö. and Pöppelbuß , J. ( 2018 ), “ Omnichannel retailing research – state of the art and intellectual foundation ”, International Journal of Physical Distribution and Logistics Management , Vol.  48 No.  4 , pp.  365 - 390 .

Gu , J. , Goetschalckx , M. and McGinnis , L.F. ( 2007 ), “ Research on warehouse operation: a comprehensive review ”, European Journal of Operational Research , Vol.  177 No.  1 , pp.  1 - 21 .

Hassan , M. , Ali , M. , Aktas , E. and Alkayid , K. ( 2015 ), “ Factors affecting selection decision of auto-identification technology in warehouse management: an international Delphi study ”, Production Planning and Control , Vol.  26 No.  12 , pp.  1025 - 1049 .

Hübner , A. , Hense , J. and Dethlefs , C. ( 2022 ), “ The revival of retail stores via omnichannel operations: a literature review and research framework ”, European Journal of Operational Research , Vol.  302 No.  3 , pp. 799 - 818 , doi: 10.1016/j.ejor.2021.12.021 .

Issaoui , Y. , Khiat , A. , Bahnasse , A. and Ouajji , H. ( 2021 ), “ Toward smart logistics: engineering insights and emerging trends ”, Archives of Computational Methods in Engineering , Vol.  28 , pp.  3183 - 3210 .

Jabbar , S. , Khan , M. , Nathalia Silva , B. and Han , K. ( 2018 ), “ A REST-based industrial web of things' framework for smart warehousing ”, Journal of Supercomputing , Vol.  74 , pp.  4419 - 4433 .

Jiang , Z.-Z. , Wan , M. , Pei , Z. and Qin , X. ( 2021 ), “ Spatial and temporal optimization for smart warehouses with fast turnover ”, Computers and Operations Research , Vol.  125 , 105091 , pp.  1 - 17 .

Kamali , A. ( 2019 ), “ Smart warehouse vs. traditional warehouse – review ”, CiiT International Journal of Automation and Autonomous System , Vol.  11 No.  1 , pp.  9 - 16 .

Kembro , J. and Norrman , A. ( 2019 ), “ Exploring trends, implications, and challenges for logistics information systems in omnichannels: Swedish retailers' perception ”, International Journal of Retail and Distribution Management , Vol.  47 No.  4 , pp.  384 - 411 .

Kembro , J.H. and Norrman , A. ( 2020 ), “ Warehouse configuration in omnichannel retailing: a multiple case study ”, International Journal of Physical Distribution and Logistics Management , Vol.  50 No.  5 , pp.  509 - 533 .

Kembro , J.H. and Norrman , A. ( 2021 ), “ Which future path to pick? A contingency approach to omnichannel warehouse configuration ”, International Journal of Physical Distribution and Logistics Management , Vol.  51 No.  1 , pp.  48 - 75 .

Kembro , J. , Danielsson , V. and Smajli , G. ( 2017 ), “ Network video technology: exploring an innovative approach to improving warehouse operations ”, International Journal of Physical Distribution and Logistics Management , Vol.  47 No.  7 , pp.  623 - 645 .

Kembro , J.H. , Norrman , A. and Eriksson , E. ( 2018 ), “ Adapting warehouse operations and design to omnichannel logistics: a literature review and research agenda ”, International Journal of Physical Distribution and Logistics Management , Vol.  48 No.  9 , pp.  890 - 912 .

Kembro , J. , Eriksson , E. and Norrman , A. ( 2022 ), “ Sorting out the sorting in omnichannel retailing ”, Journal of Business Logistics . doi: 10.1111/jbl.12305 .

Ketokivi , M. and Choi , T. ( 2014 ), “ Renaissance of case research as a scientific method ”, Journal of Operations Management , Vol.  32 No.  5 , pp.  232 - 240 .

Lee , C.K.M. , Lv , Y. , Ng , K.K.H. , Ho , W. and Choy , K.L. ( 2018 ), “ Design and application of Internet of things-based warehouse management system for smart logistics ”, International Journal of Production Research , Vol.  5 No.  8 , pp.  2753 - 2768 .

Lindecrantz , A. ( 2020 ), Stora Guiden till Retail Vem är Vem 2020 , Hakon Media AB , Solna .

Mahroof , K. ( 2019 ), “ A human-centric perspective exploring the readiness toward smart warehousing: the case of a large retail distribution warehouse ”, International Journal of Information Management , Vol.  45 , pp.  176 - 190 .

Malhotra , M.K. and Grover , V. ( 1998 ), “ An assessment of survey research in POM: from constructs to theory ”, Journal of Operations Management , Vol.  16 , pp.  407 - 425 .

MH&L ( 2020 ), “ Warehouse automation investment up due to COVID-19 ”, available at: https://www.mhlnews.com/warehousing/article/21136484/warehouse-automation-invest-up-due-to-covid19 ( accessed 15 May 2021 ).

Miles , M.B. and Huberman , A.M. ( 1994 ), Qualitative Data Analysis: An Expanded Sourcebook , 2nd ed. , Sage Publications , Thousand Oaks, California .

Reiser , C. ( 2020 ), “ Investing in warehouse automation? Who isn't ”, available at: https://logisticsviewpoints.com/2020/07/08/warehouse-automation-investment ( accessed 15 May 2021 ).

Rouwenhorst , B. , Reuter , B. , Stockrahm , V. , van Houtum , G.J. , Mantel , R.J. and Zijm , W.H. ( 2000 ), “ Warehouse design and control: framework and literature review ”, European Journal of Operational Research , Vol.  122 No.  3 , pp.  515 - 533 .

Rowley , J. (Ed.) ( 2000 ), The Principles of Warehouse Design , 2nd ed., The Institute of Logistics & Transport , Corby .

Taboada , I. and Shee , H. ( 2021 ), “ Understanding 5G technology for future supply chain management ”, International Journal of Logistics Research and Applications , Vol.  24 No.  4 , pp. 392 - 406 , doi: 10.1080/13675567.2020.1762850 .

Themistocleous , M. , Irani , Z. and O'Keefe , R.M. ( 2001 ), “ ERP and application integration. Exploratory survey ”, Business Process Management Journal , Vol.  7 No.  3 , pp.  195 - 204 .

Tokar , T. , Williams , B.D. and Fugate , B.S. ( 2020 ), “ I heart logistics – just don't ask me to pay for it: online shopper behavior in response to a delivery carrier upgrade and subsequent shipping charge increase ”, Journal of Business Logistics , Vol.  41 No.  3 , pp.  182 - 205 .

van Geest , M. , Tekinerdogan , B. and Catal , C. ( 2021 ), “ Design of a reference architecture for developing smart warehouses in Industry 4.0 ”, Computers in Industry , Vol.  124 , 103343 , pp.  1 - 21 .

Winkelhaus , S. and Grosse , E.H. ( 2020 ), “ Logistics 4.0: a systematic review toward a new logistics system ”, International Journal of Production Research , Vol.  58 No.  1 , pp.  18 - 43 .

Zhang , D. , Pee , L.G. and Cui , L. ( 2021 ), “ Artificial intelligence in e-commerce fulfillment: a case study of resource orchestration at Alibaba's smart warehouse ”, International Journal of Information Management , Vol.  57 , 102304 .

Corresponding author

Related articles, all feedback is valuable.

Please share your general feedback

Report an issue or find answers to frequently asked questions

Contact Customer Support

• SUGGESTED TOPICS
• The Magazine
• Newsletters
• Managing Yourself
• Managing Teams
• Work-life Balance
• The Big Idea
• Data & Visuals
• Reading Lists
• Case Selections
• HBR Learning
• Topic Feeds
• Account Settings
• Email Preferences

Research: Warehouse and Logistics Automation Works Better with Human Partners

• René de Koster

A recent study suggests that blending human labor with robotics leads to greater efficiency.

A study of automation usage in warehouse and logistics companies around the world suggests that blending human labor with robotics leads to greater efficiency than full automation alone. While scalable robotic systems can handle up to 1,000 tasks per hour, they often face limitations where additional robots don’t improve performance. Human-robot collaboration, employed by companies like DHL and CEVA, enhances productivity, reduces worker fatigue, and increases job satisfaction. The incremental approach of integrating human roles with automated systems not only keeps operations cost effective but also leverages human adaptability for continuous improvements.

In every sphere of business, the use of automation is growing. In warehouses and distribution, for instance, the worldwide market revenue for robotics automation is projected to grow from $7.91 billion in 2021 to more than $51 billion by 2030, according to one Statista forecast .

• RK René de Koster is a professor of logistics and operations management at Rotterdam School of Management, Erasmus University.
• DR Debjit Roy is an institute chair professor in the operations and decision sciences area at the Indian Institute of Management Ahmedabad, India.

Partner Center

• Share full article

A 2022 Supreme Court opinion.

Supported by

The Gun Lobby’s Hidden Hand in the 2nd Amendment Battle

In the battle to dismantle gun restrictions, raging in America’s courts even as mass shootings become commonplace, one name keeps turning up in the legal briefs and judges’ rulings: William English, Ph.D.

A little-known political economist at Georgetown University, Dr. English conducted a largest-of-its-kind national survey that found gun owners frequently used their weapons for self-defense. That finding has been deployed by gun rights activists to notch legal victories with far-reaching consequences.

He has been cited in a landmark Supreme Court case that invalidated many restrictions on guns, and in scores of lawsuits around the country to overturn limits on assault weapons, high-capacity magazines and the carrying of firearms. His findings were also offered in another Supreme Court case this term, with a decision expected this month.

Dr. English seems at first glance to be an impartial researcher interested in data-driven insights. He has said his “scholarly arc” focuses on good public policy, and his lack of apparent ties to the gun lobby has lent credibility to his work.

But Dr. English’s interest in firearms is more than academic: He has received tens of thousands of dollars as a paid expert for gun rights advocates, and his survey work, which he says was part of a book project, originated as research for a National Rifle Association-backed lawsuit, The New York Times has found.

He has also increasingly drawn scrutiny in some courts over the reliability and integrity of his unpublished survey, which is the core of his research, and his refusal to disclose who paid for it. Other researchers say that the wording of some questions could elicit answers overstating defensive gun use, and that he cherry-picked pro-gun responses.

The Bruen decision in 2022 upended Second Amendment law by sweeping away any modern-day gun restrictions that could not be tied to a historical antecedent. The ruling led to a surge in firearms cases — to an annual average of 680 today compared with 122 in the decade before. Pro-gun rulings have also risen: The 74 issued last year make up a quarter of all such rulings since 2000, according to researchers at the University of Southern California. Courts have struck down restrictions on high-capacity magazines in Oregon, handgun purchases in Maryland and assault weapons in California.

Dr. English’s brief in the Bruen case.

Here’s an example of that missing context.

The paper quotes a survey question, omitting the setup to it, which is highlighted below in blue.

Many policymakers recognize that a large number of people participate in shooting sports but question how often guns are used for self-defense. Have you ever defended yourself or your property with a firearm, even if it was not fired or displayed? Please do not include military service, police work, or work as a security guard.

Other questions followed the same pattern of omission. This one, about AR-15-style rifles, included text before and after the question in the version respondents saw, but not in the paper.

Some have argued that few gun owners actually want or use guns that are commonly classified as ‘assault weapons.’ Have you ever owned an AR-15 or similarly styled rifle? You can include any rifles of this style that have been modified or moved to be compliant with local law. Answering this will help us establish how popular these types of firearms are.

We are having trouble retrieving the article content.

Please enable JavaScript in your browser settings.

Thank you for your patience while we verify access. If you are in Reader mode please exit and  log into  your Times account, or  subscribe  for all of The Times.

Thank you for your patience while we verify access.

Already a subscriber?  Log in .

Want all of The Times?  Subscribe .

Advertisement

Trump handed plan to halt US military aid to Kyiv unless it talks peace with Moscow

WASHINGTON: Two key advisers to Donald Trump have presented him with a plan to end Russia’s war in Ukraine - if he wins the presidential election - that involves telling Ukraine it will only get more US weapons if it enters into peace talks. The United States would at the same time warn Moscow that any refusal to negotiate would result in increased US support for Ukraine, retired Lieutenant General Keith Kellogg, one of Trump’s national security advisers, said in an interview. Under the plan drawn up by Kellogg and Fred Fleitz, who both served as chiefs of staff in Trump’s National Security Council during his 2017-2021 presidency, there would be a ceasefire based on prevailing battle lines during peace talks. They have presented their strategy to Trump, and the former president responded favorably, Fleitz said. “I’m not claiming he agreed with it or agreed with every word of it, but we were pleased to get the feedback we did,” he said. However, Trump spokesperson Steven Cheung said only statements made by Trump or authorized members of his campaign should be deemed official. The strategy outlined by Kellogg and Fleitz is the most detailed plan yet by associates of Trump, who has said he could quickly settle the war in Ukraine if he beats President Joe Biden in the Nov. 5 election, though he has not discussed specifics. The proposal would mark a big shift in the US position on the war and would face opposition from European allies and within Trump’s own Republican Party. The Kremlin said that any peace plan proposed by a possible future Trump administration would have to reflect the reality on the ground but that Russian President Vladimir Putin remained open to talks. “The value of any plan lies in the nuances and in taking into account the real state of affairs on the ground,” Kremlin spokesman Dmitry Peskov told Reuters. “President Putin has repeatedly said that Russia has been and remains open to negotiations, taking into account the real state of affairs on the ground,” he said. “We remain open to negotiations.” Ukraine’s foreign ministry did not respond to requests for comment on the plan. NATO MEMBERSHIP ON HOLD The core elements of the plan were outlined in a publicly available research paper published by the “America First Policy Institute,” a Trump-friendly think tank where Kellogg and Fleitz hold leadership positions. Kellogg said it would be crucial to get Russia and Ukraine to the negotiating table quickly if Trump wins the election. “We tell the Ukrainians, ‘You’ve got to come to the table, and if you don’t come to the table, support from the United States will dry up,’” he said. “And you tell Putin, ‘He’s got to come to the table and if you don’t come to the table, then we’ll give Ukrainians everything they need to kill you in the field.’” According to their research paper, Moscow would also be coaxed to the table with the promise of NATO membership for Ukraine being put off for an extended period. Russia invaded neighboring Ukraine in February 2022. Until some gains by Russia in recent months, the front lines barely moved since the end of that year, despite tens of thousands of dead on both sides in relentless trench warfare, the bloodiest fighting in Europe since World War Two. Fleitz said Ukraine need not formally cede territory to Russia under their plan. Still, he said, Ukraine was unlikely to regain effective control of all its territory in the near term. “Our concern is that this has become a war of attrition that’s going to kill a whole generation of young men,” he said. A lasting peace in Ukraine would require additional security guarantees for Ukraine, Kellogg and Fleitz said. Fleitz added that “arming Ukraine to the teeth” was likely to be a key element of that. “President Trump has repeatedly stated that a top priority in his second term will be to quickly negotiate an end to the Russia-Ukraine war,” Trump spokesperson Cheung said. “The war between Russia and Ukraine never would have happened if Donald J. Trump were president. So sad.” The Biden campaign said Trump is not interested in standing up to Putin. “Donald Trump heaps praise on Vladimir Putin every chance he gets, and he’s made clear he won’t stand against Putin or stand up for democracy,” campaign spokesperson James Singer said. UPPER HAND Some Republicans will be reticent to pay for more resources to Ukraine under the plan. The US has spent more than $70 billion on military aid for Ukraine since Moscow’s invasion. “What (Trump’s supporters) want to do is reduce aid, if not turn off the spigot,” said Charles Kupchan, a senior fellow at the Council of Foreign Relations. Putin said this month that the war could end if Ukraine agreed to drop its ambitions to join NATO and hand over four eastern and southern provinces claimed by Russia. During a meeting of the United Nations Security Council last week, French and British ambassadors reiterated their view that peace can only be sought when Russia withdraws from Ukrainian territory, a position Kyiv shares. Several analysts also expressed concern that the plan by Kellogg and Fleitz could give Moscow the upper hand in talks. “What Kellogg is describing is a process slanted toward Ukraine giving up all of the territory that Russia now occupies,” said Daniel Fried, a former assistant secretary of state who worked on Russia policy. During a podcast interview last week, Trump ruled out committing US troops to Ukraine and appeared skeptical of making Ukraine a NATO member. He has indicated he would quickly move to cut aid to Kyiv if elected. Biden has consistently pushed for more Ukraine aid, and his administration supports its eventual ascension to NATO. Earlier in June, Biden and Ukrainian President Volodymyr Zelenskiy signed a 10-year bilateral security agreement .—Reuters

More from the page

India’s lawmakers take oath without duo behind bars, wikileaks’ assange set to be freed after us espionage charge plea deal, police fire on protesters trying to storm kenya parliament, several dead, more uk police officers accused in election betting scandal, germany’s covestro unveils cost cuts amid uae energy firm takeover talks.

• DOI: 10.1016/j.heliyon.2024.e33027
• Corpus ID: 270541799

Research on narrative design of handicraft intangible cultural heritage creative products based on AHP-TOPSIS method

• Min Li , Lizhe Wang , Lan Li
• Published in Heliyon 1 June 2024
• Art, Engineering

28 References

Identifying key elements for user satisfaction of bike-sharing systems: a combination of direct and indirect evaluations, on the sustainability of local cultural heritage based on the landscape narrative: a case study of historic site of qing yan yuan, china, effects of the entropy weight on topsis, research on the design of cultural and creative products in southern fujian based on narrative design methods, revision of sustainable road rating systems: selection of the best suited system for hungarian road construction using topsis method, questionnaire measures and physiological correlates of presence: a systematic review, embodied engagement with narrative: a design framework for presenting cultural heritage artifacts, raven, p g & elahi, s (2015). "the new narrative: applying narratology to the shaping of futures outputs.", applying local culture features into creative craft products design, virtually preserving the intangible heritage of artistic handicraft, related papers.

Showing 1 through 3 of 0 Related Papers

Suggestions or feedback?

MIT News | Massachusetts Institute of Technology

• Machine learning
• Social justice
• Black holes
• Classes and programs

Departments

• Aeronautics and Astronautics
• Brain and Cognitive Sciences
• Architecture
• Political Science
• Mechanical Engineering

Centers, Labs, & Programs

• Abdul Latif Jameel Poverty Action Lab (J-PAL)
• Picower Institute for Learning and Memory
• Lincoln Laboratory
• School of Architecture + Planning
• School of Engineering
• School of Humanities, Arts, and Social Sciences
• Sloan School of Management
• School of Science
• MIT Schwarzman College of Computing

Technique improves the reasoning capabilities of large language models

Press contact :, media download.

*Terms of Use:

Images for download on the MIT News office website are made available to non-commercial entities, press and the general public under a Creative Commons Attribution Non-Commercial No Derivatives license . You may not alter the images provided, other than to crop them to size. A credit line must be used when reproducing images; if one is not provided below, credit the images to "MIT."

Previous image Next image

Large language models like those that power ChatGPT have shown impressive performance on tasks like drafting legal briefs, analyzing the sentiment of customer reviews, or translating documents into different languages.

These machine-learning models typically use only natural language to process information and answer queries, which can make it difficult for them to perform tasks that require numerical or symbolic reasoning.

For instance, a large language model might be able to memorize and recite a list of recent U.S. presidents and their birthdays, but that same model could fail if asked the question “Which U.S. presidents elected after 1950 were born on a Wednesday?” (The answer is Jimmy Carter.)

Researchers from MIT and elsewhere have proposed a new technique that enables large language models to solve natural language, math and data analysis, and symbolic reasoning tasks by generating programs.

Their approach, called natural language embedded programs (NLEPs), involves prompting a language model to create and execute a Python program to solve a user’s query, and then output the solution as natural language.

They found that NLEPs enabled large language models to achieve higher accuracy on a wide range of reasoning tasks. The approach is also generalizable, which means one NLEP prompt can be reused for multiple tasks.

NLEPs also improve transparency, since a user could check the program to see exactly how the model reasoned about the query and fix the program if the model gave a wrong answer.

“We want AI to perform complex reasoning in a way that is transparent and trustworthy. There is still a long way to go, but we have shown that combining the capabilities of programming and natural language in large language models is a very good potential first step toward a future where people can fully understand and trust what is going on inside their AI model,” says Hongyin Luo PhD ’22, an MIT postdoc and co-lead author of a paper on NLEPs .

Luo is joined on the paper by co-lead authors Tianhua Zhang, a graduate student at the Chinese University of Hong Kong; and Jiaxin Ge, an undergraduate at Peking University; Yoon Kim, an assistant professor in MIT’s Department of Electrical Engineering and Computer Science and a member of the Computer Science and Artificial Intelligence Laboratory (CSAIL); senior author James Glass, senior research scientist and head of the Spoken Language Systems Group in CSAIL; and others. The research will be presented at the Annual Conference of the North American Chapter of the Association for Computational Linguistics.

Problem-solving with programs

Many popular large language models work by predicting the next word, or token, given some natural language input. While models like GPT-4 can be used to write programs, they embed those programs within natural language, which can lead to errors in the program reasoning or results.

With NLEPs, the MIT researchers took the opposite approach. They prompt the model to generate a step-by-step program entirely in Python code, and then embed the necessary natural language inside the program.

An NLEP is a problem-solving template with four steps. First, the model calls the necessary packages, or functions, it will need to solve the task. Step two involves importing natural language representations of the knowledge the task requires (like a list of U.S. presidents’ birthdays). For step three, the model implements a function that calculates the answer. And for the final step, the model outputs the result as a line of natural language with an automatic data visualization, if needed.

“It is like a digital calculator that always gives you the correct computation result as long as the program is correct,” Luo says.

The user can easily investigate the program and fix any errors in the code directly rather than needing to rerun the entire model to troubleshoot.

The approach also offers greater efficiency than some other methods. If a user has many similar questions, they can generate one core program and then replace certain variables without needing to run the model repeatedly.

To prompt the model to generate an NLEP, the researchers give it an overall instruction to write a Python program, provide two NLEP examples (one with math and one with natural language), and one test question.

“Usually, when people do this kind of few-shot prompting, they still have to design prompts for every task. We found that we can have one prompt for many tasks because it is not a prompt that teaches LLMs to solve one problem, but a prompt that teaches LLMs to solve many problems by writing a program,” says Luo.

“Having language models reason with code unlocks many opportunities for tool use, output validation, more structured understanding into model's capabilities and way of thinking, and more,” says Leonid Karlinsky, principal scientist at the MIT-IBM Watson AI Lab.

“No magic here”

NLEPs achieved greater than 90 percent accuracy when prompting GPT-4 to solve a range of symbolic reasoning tasks, like tracking shuffled objects or playing a game of 24, as well as instruction-following and text classification tasks. The researchers found that NLEPs even exhibited 30 percent greater accuracy than task-specific prompting methods. The method also showed improvements over open-source LLMs. 

Along with boosting the accuracy of large language models, NLEPs could also improve data privacy. Since NLEP programs are run locally, sensitive user data do not need to be sent to a company like OpenAI or Google to be processed by a model.

In addition, NLEPs can enable small language models to perform better without the need to retrain a model for a certain task, which can be a costly process.

“There is no magic here. We do not have a more expensive or fancy language model. All we do is use program generation instead of natural language generation, and we can make it perform significantly better,” Luo says.

However, an NLEP relies on the program generation capability of the model, so the technique does not work as well for smaller models which have been trained on limited datasets. In the future, the researchers plan to study methods that could make smaller language models generate more effective NLEPs. In addition, they want to investigate the impact of prompt variations on NLEPs to enhance the robustness of the model’s reasoning processes.

This research was supported, in part, by the Center for Perceptual and Interactive Intelligence of Hong Kong. 

Share this news article on:

Related links.

• Hongyin Luo
• Computer Science and Artificial Intelligence Laboratory (CSAIL)
• Department of Electrical Engineering and Computer Science

Related Topics

• Computer science and technology
• Artificial intelligence
• Programming
• Electrical Engineering & Computer Science (eecs)

Related Articles

Natural language boosts LLM performance in coding, planning, and robotics

Large language models use a surprisingly simple mechanism to retrieve some stored knowledge

Explained: Generative AI

Learning to grow machine-learning models

Previous item Next item

More MIT News

A home away from a homeland

Read full story →

“UnrulyArt” creates joy and engagement, regardless of ability

Owen Coté, military technology expert and longtime associate director of the Security Studies Program, dies at 63

What happens during the first moments of butterfly scale formation

Startup aims to transform the power grid with superconducting transmission lines

Designing for outer space

• More news on MIT News homepage →

Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA, USA

• Map (opens in new window)
• Events (opens in new window)
• People (opens in new window)
• Careers (opens in new window)
• Accessibility
• Social Media Hub
• MIT on Facebook
• MIT on YouTube
• MIT on Instagram

Cryomicrophotography in a Cold Room at the Faculty of Geography of Moscow State University Named After M.V. Lomonosov

• Conference paper
• First Online: 26 June 2024
• Cite this conference paper

• D. M. Frolov   ORCID: orcid.org/0000-0002-0307-8175 5 ,
• G. A. Rzhanitsyn 5 ,
• A. V. Koshurnikov   ORCID: orcid.org/0000-0001-6160-7795 5 ,
• V. E. Gagarin   ORCID: orcid.org/0000-0001-8100-4000 5 &
• P. A. Krilovets 5  

Part of the book series: Springer Proceedings in Earth and Environmental Sciences ((SPEES))

Included in the following conference series:

• Conference on Physical and Mathematical Modeling of Earth and Environment Processes

At the Faculty of Geography of Moscow State University named after M.V. Lomonosov, a cold room was equipped to study the properties of snow, ice and frozen soil for the needs of the educational process and scientific research at the Department of Cryolithology and Glaciology and the Laboratory of Snow Avalanches and Mudflows of the Faculty of Geography. In particular, the cold laboratory is equipped for optical photography under a microscope and in polarized light.

This is a preview of subscription content, log in via an institution to check access.

Acknowledgements

The work was carried out in accordance with the state budget theme “Evolution of the cryosphere under climate change and anthropogenic impact” (121051100164-0), “Danger and risk of natural processes and phenomena” (121051300175-4). The authors declare no conflict of interest.

Author information

Authors and affiliations.

Faculty of Geography of Moscow State University Named After M.V. Lomonosov, St. Leninskiye Gory, 1, Moscow, 119331, Russia

D. M. Frolov, G. A. Rzhanitsyn, A. V. Koshurnikov, V. E. Gagarin & P. A. Krilovets

You can also search for this author in PubMed   Google Scholar

Corresponding author

Correspondence to D. M. Frolov .

Editor information

Editors and affiliations.

Fontanars dels Alforins, Valencia, Spain

Vladimir Karev

Rights and permissions

Reprints and permissions

Copyright information

© 2024 The Author(s), under exclusive license to Springer Nature Switzerland AG

About this paper

Cite this paper.

Frolov, D.M., Rzhanitsyn, G.A., Koshurnikov, A.V., Gagarin, V.E., Krilovets, P.A. (2024). Cryomicrophotography in a Cold Room at the Faculty of Geography of Moscow State University Named After M.V. Lomonosov. In: Karev, V. (eds) Proceedings of the 9th International Conference on Physical and Mathematical Modelling of Earth and Environmental Processes. PMMEEP 2023. Springer Proceedings in Earth and Environmental Sciences. Springer, Cham. https://doi.org/10.1007/978-3-031-54589-4_9

Download citation

DOI : https://doi.org/10.1007/978-3-031-54589-4_9

Published : 26 June 2024

Publisher Name : Springer, Cham

Print ISBN : 978-3-031-54588-7

Online ISBN : 978-3-031-54589-4

eBook Packages : Earth and Environmental Science Earth and Environmental Science (R0)

Share this paper

Anyone you share the following link with will be able to read this content:

Sorry, a shareable link is not currently available for this article.

Provided by the Springer Nature SharedIt content-sharing initiative

• Publish with us

Policies and ethics

• Find a journal
• Track your research