Author: Evelyn Yao, undergraduate student at Tsinghua University
editor's note
In the "Selected Recent Papers of Operations Research", we have selected some interesting articles in Operations Research with themes and pertinence, not only summarizing and commenting on the content of the articles, but also sorting out the structure of the articles, It aims to stimulate readers' interest in reading and enthusiasm for exploration. In this issue of "Selected Papers", we take "the future of retail industry" as the theme, and explore the path planning problem in the delivery of self-driving unmanned retail vehicles, the slowdown delay problem in online retailing, and perishable items. Inventory management problems, involving queuing systems, statistics, Markov decision process and many other knowledge.
Recommended Article 1
Release time: 2022.12.1
● 题目：Stall Economy: The Value of Mobility in Retail on Wheels
The Stall Economy: The Value of Retail Cart Mobility
● Journal : Operations Research
● Original link : https://doi.org/10.1287/opre.2022.2404
● Authors : Junyu Cao, Wei Qi
● Keywords : stall economy (stall economy) • mobile retail (mobile retail) • self-driving (autonomous driving) • rendezvous problem (convergence problem) • spatial queues (space queuing)
● Summary :
Urban open space emerges as a new territory to embrace retail innovations. Selling products in public spaces with wheeled stalls can potentially become ubiquitous in our future cities. Transition into such a “stall economy” paradigm is being spurred by the rapidly advancing self-driving technologies. Motivated by this transformation, this paper provides models, theory, and insights of spatial queueing systems, in which one server moves around to meet mobile customers/machines and in which the “last 100 meters” are expensive. Specifically, we study two service modes: (i) on-demand, first come, first served and (ii) spatially and temporally pooling customer demands. In each mode, we derive the dependence of customer waiting and stall repositioning on two key decisions: the service zone size and the walking distance imposed on customers to meet a stall. In particular, for the on-demand mode, we propose and solve a “rendezvous problem” to analytically characterize the spatial distribution of the stall-customer meeting locations. We also propose a stylized joint truck-stall routing model to capture the inventory replenishment operations. Our main finding is that the stall economy potentially profits more than stationary retail, not only because of the mobility of stalls for providing proximity to customers, but also because of its operational flexibilities that allow for avoiding the “last 100 meters” and pooling demands. In a broader sense, this work looks toward an expanded scope of future retail empowered by self-driving technologies.
Urban open space is gradually becoming a new field for embracing retail innovation. The use of retail vans to sell products in urban public spaces is also likely to become a common phenomenon in our future cities. Motivated by our transition to a "stall economy" driven by rapidly advancing autonomous driving technologies, this paper provides models, theory, and insights into spatial queuing systems for the "last 100 meters" problem of mobile retail vehicles. Specifically, we study two service models: (i) on-demand, first-come, first-served; (ii) aggregation of customer demands in space and time. In each model, we derive the dependence of customer waiting time and retail cart repositioning costs on the size of the service area and the distance traveled by customers. Especially for the on-demand service model, we propose and solve a "confluence problem" to analyze the spatial distribution characteristics of retail carts and customers' confluence locations. We also propose a joint model of stock carts and retail carts to capture stock replenishment operations. Our main finding is that the stall economy has the potential to be more profitable than stationary retailing, not only because of the fluidity of the stall, which provides close proximity to customers, but also because of its operational flexibility, which avoids " Expensive expense for the last 100 meters". In a broader sense, this work looks at expanding the scope of future retail empowered by autonomous driving technology.
● Article structure :
● Comments :
This article studies the path planning problem during the "last mile" delivery of unmanned retail vehicles, and the topic selection is particularly novel. In the core part of the article, the author proposes a very beautiful magnet rule and uses it to obtain a specific confluence point distribution. In section 5.1 of the article, the magnet rule is also compared with other rules (such as straight-ahead rule and two-location clairvoyant, etc.), and it is a very efficient rule after comprehensive consideration. During the communication between the editor and the author, the author mentioned that in his opinion, the proposal of the magnet rule is the most creative part of the whole article. Not only that, but the author also hopes that his article can bring some inspiration and inspiration to everyone. New ideas, applying theory together to improve efficiency, and in turn promoting theoretical innovation through the exploration of real-world problems.
Recommended Article 2
Release time: 2022.3.14
● 题目：Need for Speed: The Impact of In-Process Delays on Customer Behavior in Online Retail
In pursuit of speed: The impact of delays in the process on customer behavior
● Journal : Operations Research
● Original link : https://doi.org/10.1287/opre.2022.2262
● Footnote : James Gallino, Nil Karacaoglu, Antonio Moreno
● Keywords : in-process delays (delay in the process) • service operations (service operations) • online retail (online retail) • website speed (website speed) • abandonment (abstainment)
● Summary :
The impact of delays has been widely studied in various offline services. The focus of this study is online services, and we explore the impact of in-process delays—measured by website speed—on customer behavior. We leverage novel retail and website speed data to investigate how delays impact online sales and how customer sensitivity to in-process delays varies across the different stages of a customer’s shopping journey. We estimate sizable adverse effects of website slowdowns on online sales. Using threshold regression models, we show that customers exhibit diminishing sensitivity to increases in website slowdowns. Our results suggest that waiting times affect customer abandonment differently at different stages of the shopping journey. Customers are more sensitive to slowdowns at the checkout stage. Our findings have implications for website design decisions such as improving website speed at the checkout stage, selecting third-party content providers, and customizing the design of mobile and desktop channels. The paper’s results are especially relevant in the current regulatory environment with ongoing policy debates about net neutrality.
We have extensively studied the impact of latency in various offline services. Focusing on online services, this study explores the impact of "in-process" delays (measured by website speed) on customer behavior. We use speed data from new retail websites to explore how "latency" affects online sales and how shoppers' sensitivity to network latency changes at different stages of their shopping. Our research estimates that a slowdown in network speeds has a considerable detrimental effect on online sales. Using a threshold regression model, we show that customers exhibit decreasing sensitivity to increases in website latency. Our results show that the impact of waiting time on customer abandonment is different at different stages of the shopping journey, and customers are more sensitive to network delays at the checkout stage. Our findings have implications for site design decisions, such as improving site speed at the checkout stage, selecting third-party content providers, and customizing designs for mobile and desktop channels. Especially in the current regulatory environment where net neutrality policy disputes continue, our research is even more meaningful.
● Article structure :
● Comments :
This paper focuses on the relationship between "the delay time of the website and the customer transaction in online shopping". Compared with the complex derivation and beautiful and wonderful rules in the previous article, this article is more friendly to novices. As long as you have basic statistical knowledge, you can fully understand the connotation of the article. In terms of topic selection, the article discovers the gap in the previous literature, and realizes some supplements to the related literature on "the impact of delay on online retail services". It is very suitable for eating (×) and expanding knowledge.
Recommended Article 3
Release time: 2022.12.23
● 题目：Inventory Sharing for Perishable Products: Application to Platelet Inventory Management in Hospital Blood Banks
Inventory Sharing for Perishable Products: Platelet Inventory Management in Hospital Blood Banks
● Journal : Operations Research
● Original link : https://doi.org/10.1287/opre.2022.2410
● 作者：Can Zhang, Turgay Ayer, Chelsea C. White, Joy N. Bodeker， John D. Roback
● Keywords : blood banks (blood bank) • platelets (platelets) • perishable products (perishable products) • transshipment (transshipment) • value of inventory sharing (value of inventory sharing)
● Summary :
Platelets are critical blood products. The management of platelet inventory is particularly challenging because of its perishable nature with a short shelf life. Motivated by a platelet inventory management problem at a two-location hospital system, we study how the wastage of platelets and, more broadly, perishable products can be reduced through inventory sharing. In particular, we consider a system with two locations and a single product (e.g., a two-hospital system sharing blood products, such as platelets). Each location faces a stochastic demand, and products can be transshipped from one location to the other after demand realization. At each location, products are issued in a first-in, first-out manner. Although the state of such a complex system consists of the inventory levels of different product ages at both locations, interestingly, we show that the direction of transshipment can be determined by simply comparing the age of the oldest products at each location after meeting demand. Based on this and other structural results, we then prove that a myopic transshipment policy is optimal for a special case motivated by our case study and serves as a lower bound on the optimal transshipment quantity for more general settings. Our analysis also sheds light on how inventory sharing affects the optimal inventory levels for perishable products. Because of its competitive performance and simplicity, our proposed myopic transshipment policy has been implemented by our partner hospital system, which led to a reduction of approximately 20% in platelet outdates.
Platelets are important blood products. However, due to the short shelf life and fragile nature of platelets, his inventory management was particularly challenging. Inspired by the problem of platelet inventory management in a bipartite hospital system, we investigate how inventory sharing can reduce the waste of platelets and, more generally, perishable products. In particular, we consider a system with two sites and a single product (eg, a system consisting of two hospitals that are able to share blood products such as platelets). The demand faced by each location is random, and the product can be transferred from one location to another after the demand is fulfilled. At each location, products are issued on a first-in, first-out basis. Although the state of such a complex system consists of inventory levels with differing product ages in two locations, interestingly we show that the direction of transshipment can be determined simply by comparing the age of the oldest product at each location after satisfying demand. Based on this and other structural results, we then demonstrate that the transshipment policy of myopia is optimal in a special case motivated by our case study and can serve as a lower bound on the optimal number of transshipments in the more general case. Our analysis also reveals how inventory sharing affects optimal stock levels for perishable products. Due to the simplicity and competitiveness of our proposed protocol, our proposed myopia transfer policy has now been implemented by our partner hospital systems, and the platelet expiration rate has been reduced by about 20% after implementation.
● Article structure :
● Comments :
This paper explores the problem of inventory sharing of perishable items in the two systems and provides an implementable transshipment strategy. Not only that, some insights and discoveries in this paper are also of great practical significance. For example, when storing perishable products, transshipment should happen more frequently, etc. In terms of research methods, this paper is based on stochastic process modeling, and thus deduces the properties of the optimal strategy, such as the transport direction and the upper and lower limits of the optimal transport quantity. If you are interested in dynamic inventory management, please read this article.