编者按：
在本系列文章中，我们梳理了运筹学顶刊MSOM在2024年6月份发布的7篇文章的基本信息，旨在帮助读者快速洞察行业最新动态。
文章1
题目
Promotional Inventory Displays: An Empirical Analysis Using IoT Data促销库存展示：使用物联网数据的实证分析
作者
Jacob Z. Zeng, Ashish Agarwal, Ioannis Stamatopoulos
原文链接
https://doi.org/10.1287/msom.2022.0291
发布时间
4 Jun 2024
摘要
Problem definition: Despite their widespread use, promotional inventory displays’ campaign execution is opaque. Brands (manufacturers) can only verify in-store display presence through manual, on-site audits, which are costly and limited in scope. This lack of visibility into execution makes it difficult for brands to quantify the sales impact of displays and properly evaluate campaign performance.
Methodology/results: We use Internet-of-Things (IoT)-generated data on the real-time location of about 15,000 displays from 10 display campaigns in about 5,000 stores of a Fortune 500 retail chain, paired with the stores’ point-of-sale (POS) data between September 2017 and March 2018, to measure the operational execution and effectiveness of display campaigns. First, we find that campaigns are poorly executed: 29% of displays never made it to a store’s floor, and those that made it only spent 62% of the campaign there. Second, we find that poor execution deprives brands of substantial sales, especially during campaign weeks: placing a display on the floor during an arbitrary week increases the targeted products’ sales by 7.3%, and placing it on the floor during a campaign week boosts sales by another 2.3%.
Managerial implications: Our results enable brands to optimize investments in promotional display campaigns and execution. We project that brands would suffer a 3.1% dollar sales decrease during campaign weeks by discontinuing current campaigns. However, they could achieve up to a 6.9% dollar sales increase during campaign weeks through improved execution.
问题定义：尽管促销库存展示被广泛使用，但其活动执行过程是不透明的。品牌（制造商）只能通过成本高昂且范围有限的现场手动审核来验证店内展示的存在。这种对执行情况的不可见性使得品牌难以量化展示对销售的影响，以及正确评估活动表现。
方法/结果：我们使用物联网（IoT）生成的数据，这些数据涉及大约5000家财富500强零售连锁店中约15000个展示的实时位置，以及这些商店的销售点（POS）数据，时间跨度从2017年9月到2018年3月，来衡量展示活动的运营执行和效果。首先，我们发现活动执行情况很差：29%的展示从未到达商店的地面，而那些到达的展示在活动期间只有62%的时间在商店。其次，我们发现执行不佳剥夺了品牌大量的销售，特别是在活动周：在任意一周将展示放置在地面上，目标产品的销售增加了7.3%，在活动周放置在地面上则销售又增加了2.3%。
管理意义：我们的结果使品牌能够优化在促销展示活动和执行上的投资。我们预计，如果停止当前的活动，品牌在活动周的销售额将下降3.1%。然而，通过改进执行，他们可以在活动周实现高达6.9%的销售额增长。
文章2
题目
Dual Value of Delayed Incentives: An Empirical Investigation of Gift Card Promotions延迟激励的双重价值：对礼品卡促销的实证研究
作者
Bharadwaj Kadiyala, Özalp Özer, A. Serdar Şimşek
原文链接
https://doi.org/10.1287/msom.2022.0218
发布时间
10 Jun 2024
摘要
Problem definition: Gift card promotions offer a tiered incentive (a gift card) to customers for purchasing regularly-priced products, with higher expenditure receiving a larger incentive. We empirically investigate the impact of gift card promotions on customer purchase behavior by collaborating with a major U.S.-based department store that targets gift card promotion emails on its online channel.
Methodology/results: We use a collection of discontinuities in the retailer’s targeting policies and corresponding fuzzy regression discontinuity designs to estimate localized causal effects of the gift card promotion. We find that gift card promotion provided to customers whose most recent purchase was 4 and 13 months ago, respectively, increased their purchase probability by 11.24% and 26.64%, and their average expenditure by 9.53% and 1.9%, respectively. Overall, the gift card promotion generated, on average, $100K in incremental sales per promotion from these customers alone. Furthermore, the promotion induced customers to seek out products from niche categories and increased spending on high-end brands. Gift card promotion emails also had a dominant advertisement effect, that is, customers who received the gift card promotion email increased expenditure even if they did not participate in the promotion.
Managerial implications: (i) Gift card promotions are a viable strategy to boost short-term sales and to alter the distribution of the types (brand, category) of purchased products. (ii) Retailers should optimize gift card promotion tiers based on the tradeoff between increased customer expenditure and promotion expenses and consider the advertisement effect when targeting the promotion emails. (iii) Contrary to popular belief, retailers benefit from customers’ redeeming gift cards because they spend more due to the gift card, validating the dual value of the delayed incentive.
问题定义：礼品卡促销活动为购买正价商品的顾客提供分层激励（礼品卡），消费越高，获得的激励越大。我们与一家主要的美国百货商店合作，通过其在线渠道针对礼品卡促销电子邮件，实证研究礼品卡促销对顾客购买行为的影响。
方法/结果：我们利用零售商目标政策中的一系列不连续性，以及相应的模糊回归不连续性设计，估计礼品卡促销的局部因果效应。我们发现，分别针对最近一次购买发生在4个月和13个月前的顾客提供的礼品卡促销，使他们的购买概率分别增加了11.24%和26.64%，平均消费分别增加了9.53%和1.9%。总体而言，仅从这些顾客那里，每次礼品卡促销平均产生了10万美元的额外销售额。此外，促销活动还促使顾客寻找小众类别的产品，并增加了对高端品牌的支出。礼品卡促销电子邮件还具有主导的广告效果，即即使顾客没有参与促销活动，收到礼品卡促销电子邮件的顾客也增加了支出。
管理意义：（i）礼品卡促销是一种可行的策略，可以促进短期销售并改变购买产品类型（品牌、类别）的分布。（ii）零售商应根据增加顾客支出和促销费用之间的权衡，优化礼品卡促销层级，并在定位促销电子邮件时考虑广告效果。（iii）与普遍看法相反，零售商从顾客兑换礼品卡中受益，因为他们由于礼品卡而增加了支出，从而验证了延迟激励的双重价值。
文章3
题目
Direct Trade Sourcing Strategies for Specialty Coffee特种咖啡的直接贸易采购策略
作者
Scott Webster, Burak Kazaz, Shahryar Gheibi
原文链接
https://doi.org/10.1287/msom.2021.0586
发布时间
12 Jun 2024
摘要
Problem definition: Leading specialty coffee roasters rely on direct trade to source premium coffee beans. We examine a roaster who sells two basic types of roasts: (1) a single-origin roast sourced from a specific locale and (2) a blend roast that uses a mix of beans from sources that vary over the course of a year. The prices of blend roasts are lower than those of single-origin roasts and appeal to a larger market. We study how characteristics of the operating and market environment affect the optimal sourcing strategy for single-origin beans.
Methodology/results: We develop a two-stage stochastic program with recourse that reflects these characteristics. A roaster has the option to allocate some of the single-origin beans for sale under a blend label, known as downward substitution. We identify three distinct optimal sourcing strategies—specialized (no downward substitution), diversified (consistent downward substitution), and mixed (between these extremes)—and show that they are robust under different definitions of yield and demand.
Managerial implications: We identify four main insights: (1) Two factors determine which strategy is optimal: the mean price of the inferior product (blend label) and the marginal cost of the superior product (single-origin label). (2) When compared with the newsvendor model, we find distinct structural differences across strategies. For example, whereas the effects of increasing uncertainty on optimal quantity align with the newsvendor model under a mixed strategy, the effects are distinctly different under specialized and diversified strategies (e.g., monotonic decreasing behavior for specialized, no change in quantity under diversified). (3) The weighted average price of an agricultural product is decreasing in negative yield–price correlation. We coin this as the “farmer’s curse,” which carries lessons for direct trade sourcing (e.g., advocating against paying the grower at postharvest market prices). (4) We find evidence of a virtuous feedback loop wherein the grower–roaster relationship becomes stronger over time. Our findings also point to a simple signal that policymakers may use to identify coffee growing locales where targeted interventions can improve grower welfare.
问题定义：领先的特种咖啡烘焙商依赖直接贸易来采购优质咖啡豆。我们研究了一家销售两种基本类型烘焙咖啡的烘焙商：(1) 来自特定地区的单一产地烘焙咖啡，(2) 使用一年中不同来源混合豆的混合烘焙咖啡。混合烘焙咖啡的价格低于单一产地烘焙咖啡，并且吸引了更大的市场。我们研究了运营和市场环境的特征如何影响单一产地豆的最佳采购策略。
方法/结果：我们开发了一个具有追索权的两阶段随机规划模型，以反映这些特征。烘焙商可以选择将一些单一产地豆以混合标签出售，这被称为向下替代。我们确定了三种不同的最佳采购策略——专业化（无向下替代）、多样化（持续向下替代）和混合（介于这两种极端之间）——并表明它们在不同的产量和需求定义下是稳健的。
管理意义：我们确定了四个主要的洞见：(1) 两个因素决定哪种策略是最佳：劣质产品（混合标签）的平均价格和优质产品（单一产地标签）的边际成本。(2) 与新闻摊贩模型相比，我们发现不同策略之间存在明显结构差异。例如，与混合策略下的新闻摊贩模型相比，增加不确定性对最佳数量的影响是一致的，而在专业化和多样化策略下的影响则截然不同（例如，专业化策略下的最佳数量呈单调递减行为，多样化策略下数量不变）。(3) 农产品的加权平均价格随着产量与价格的负相关性而降低。我们称之为“农民的诅咒”，这对直接贸易采购有启示（例如，反对以收获后市场价格向种植者支付）。(4) 我们发现了一种良性反馈循环的证据，即种植者与烘焙商之间的关系随着时间的推移变得更加牢固。我们的发现还指向了一个简单的信号，政策制定者可以使用它来识别咖啡种植地区，那里的有针对性的干预可以改善种植者的福祉。
文章4
题目
Capacity Allocation and Scheduling in Two-Stage Service Systems with Multiclass Customers多类客户的两阶段服务系统中的容量分配和调度
作者
Zhiheng Zhong, Ping Cao, Junfei Huang, Sean X. Zhou
原文链接
https://doi.org/10.1287/msom.2023.0266
发布时间
13 Jun 2024
摘要
Problem definition: This paper considers a tandem queueing system in which stage 1 has one station serving multiple classes of arriving customers with different service requirements and related delay costs, and stage 2 has multiple parallel stations, with each station providing one type of service. Each station has many statistically identical servers. The objective is to design a joint capacity allocation between stages/stations and scheduling rule of different classes of customers to minimize the system’s long-run average cost.
Methodology/results: Using fluid approximation, we convert the stochastic problem into a fluid optimization problem and develop a solution procedure. Based on the solution to the fluid optimization problem, we propose a simple and easy-to-implement capacity allocation and scheduling policy and establish its asymptotic optimality for the stochastic system. The policy has an explicit index-based scheduling rule that is independent of the arrival rates, and resource allocation is determined by the priority orders established between the classes and stations. We conduct numerical experiments to validate the accuracy of the fluid approximation and demonstrate the effectiveness of our proposed policy.
Managerial implications: Tandem queueing systems are ubiquitous. Our results provide useful guidelines for the allocation of limited resources and the scheduling of customer service in those systems. Our proposed policy can improve the system’s operational efficiency and customers’ service quality.
问题定义：本文考虑了一个串联排队系统，其中第一阶段有一个服务站为多类到达顾客提供不同服务需求和相关延迟成本的服务，第二阶段有多个并行服务站，每个服务站提供一种类型的服务。每个服务站有许多统计上相同的服务人员。目标是设计一个联合的阶段/站点容量分配和不同类别顾客的调度规则，以最小化系统的长期平均成本。
方法/结果：利用流体近似，我们将随机问题转化为流体优化问题，并开发了一种解决方案。基于流体优化问题的解，我们提出了一个简单且易于实施的容量分配和调度策略，并建立了其对随机系统的渐近最优性。该策略具有明确的基于指数的调度规则，该规则与到达率无关，资源分配由在类别和站点之间建立的优先级顺序决定。我们进行数值实验来验证流体近似的准确性，并展示我们提出的策略的有效性。
管理意义：串联排队系统无处不在。我们的结果为这些系统中有限资源的分配和顾客服务的调度提供了有用的指导。我们提出的策略可以提高系统的运行效率和顾客的服务品质。
文章5
题目
Adaptive Seamless Dose-Finding Trials自适应无缝剂量寻找试验
作者
Ningyuan Chen, Amin Khademi
原文链接
https://doi.org/10.1287/msom.2023.0246
发布时间
13 Jun 2024
摘要
Problem definition: We study early-stage dose-finding clinical trials with simultaneous consideration of efficacy and toxicity without parametric assumptions on the forms of the unknown dose-efficacy and dose-toxicity curves. We propose algorithms that adaptively allocate doses based on patient responses, in order to maximize the efficacy for the patients during the trial while minimizing the toxicity.
Methodology/results: We leverage online learning to design the clinical trial and propose two algorithms. The first one follows dose-escalation principles and analyzes the efficacy and toxicity simultaneously. The second one uses bisection search to identify a safe dose range and then applies upper confidence bound algorithms within the safe range to identify efficacious doses. We show the matching upper and lower bounds for the regret of both algorithms. We find that observing the dose-escalation principle is costly, as the optimal regret of the first algorithm is in the order of N3/4, worse than the optimal regret of the second algorithm, which is in the order of N2/3. We test our proposed algorithms with three benchmarks commonly used in practice on synthetic and real data sets, and the results show that they are competitive with or significantly outperform the benchmarks.
Managerial implications: We provide a novel insight that following the dose-escalation principle inevitably leads to higher regret. The first proposed algorithm is suitable to use when little information about the dose-toxicity profile is available, whereas the second one is appealing when more information is available about the toxicity profile.
问题定义：我们研究了同时考虑效果和毒性的早期剂量寻找临床试验，不对未知剂量-效果和剂量-毒性曲线的形式进行参数假设。我们提出了基于患者反应自适应分配剂量的算法，以便在试验期间最大化患者的疗效，同时最小化毒性。
方法/结果：我们利用在线学习来设计临床试验，并提出了两种算法。第一种遵循剂量递增原则，同时分析效果和毒性。第二种使用二分搜索来确定安全剂量范围，然后在安全范围内应用上置信界算法来识别有效剂量。我们展示了两种算法的遗憾度（regret）的匹配上下界。我们发现，遵守剂量递增原则代价较高，因为第一种算法的最优遗憾度为N的3/4次方，比第二种算法的最优遗憾度（N的2/3次方）要差。我们在合成数据和真实数据集上用三种常用于实践的基准测试了我们提出的算法，结果显示它们与基准相当或显著优于基准。
管理意义：我们提供了一个新的洞见，即遵循剂量递增原则将不可避免地导致更高的遗憾度。当对剂量-毒性概况了解甚少时，适合使用第一种提出的算法；而当对毒性概况有更多的了解时，第二种算法则更具吸引力。
文章6
题目
Routing and Staffing in Customer Service Chat Systems with Generally Distributed Service and Patience Times具有一般分布的服务时间和耐心时间的客户服务聊天系统中的路由与人员配置
作者
Zhenghua Long, Tolga Tezcan, Jiheng Zhang
原文链接
https://doi.org/10.1287/msom.2022.0114
发布时间
14 Jun 2024
摘要
Problem definition: We study customer service chat (CSC) systems, in which agents can serve multiple customers simultaneously, with generally distributed service and patience times. The multitasking capability of agents introduces idiosyncratic challenges when making routing and staffing decisions.
Methodology/results: To determine the dynamic matching of arriving customers with available agents, we first formulate a routing linear program (LP) based on system primitives. Inspired by the optimal solution of the routing LP, we design a parsimonious dynamic routing policy that is independent of arrival rate and service capacity information. We also use the optimal solution to develop closed-form approximations for crucial performance metrics and show that a similar LP can be utilized to make staffing decisions. Through extensive simulation experiments, we showcase the efficacy of our approximations and staffing decisions. Furthermore, under our proposed policy, the CSC system exhibits a unique stationary fluid model in which the steady-state performance measures align with our approximations.
Managerial implications: The extant literature primarily focuses on Markovian systems with exponential distributions. In this paper, customers’ service and patience times are allowed to be generally distributed to agree with practical settings. Our findings indicate that the distributions have a significant impact on routing policies, staffing decisions, and system performance.
问题定义：我们研究客户服务聊天（CSC）系统，其中代理可以同时为多个客户提供服务，服务时间和耐心时间具有一般分布。代理的多任务处理能力在做出路由和人员配置决策时引入了特有的挑战。
方法/结果：为了确定到达客户与可用代理的动态匹配，我们首先基于系统原语制定路由线性规划（LP）。受路由LP最优解的启发，我们设计了一个节俭的动态路由策略，该策略不依赖于到达率和服务能力信息。我们还使用最优解开发了关键性能指标的封闭形式近似，并展示了类似的LP可以用于做出人员配置决策。通过广泛的模拟实验，我们展示了我们的近似和人员配置决策的有效性。此外，在我们的提议策略下，CSC系统展现出一个独特的静态流体模型，其中稳态性能指标与我们的近似一致。
管理意义：现有文献主要关注具有指数分布的马尔可夫系统。在本文中，允许客户的服务和耐心时间具有一般分布，以符合实际设置。我们的发现表明，这些分布对路由策略、人员配置决策和系统性能有显著影响。
文章7
题目
How to Conclude a Suspended Sports League? 如何结束一个暂停的体育联赛？
作者
Ali Hassanzadeh, Mojtaba Hosseini, John G. Turner
原文链接
https://doi.org/10.1287/msom.2022.0558
发布时间
18 Jun 2024
摘要
Problem definition: Professional sports leagues may be suspended because of various reasons, such as the recent coronavirus disease 2019 pandemic. A critical question that the league must address when reopening is how to appropriately select a subset of the remaining games to conclude the season in a shortened time frame. Despite the rich literature on scheduling an entire season starting from a blank slate, concluding an existing season is quite different. Our approach attempts to achieve team rankings similar to those that would have resulted had the season been played out in full.
Methodology/results: We propose a data-driven model that exploits predictive and prescriptive analytics to produce a schedule for the remainder of the season composed of a subset of originally scheduled games. Our model introduces novel rankings-based objectives within a stochastic optimization model, whose parameters are first estimated using a predictive model. We introduce a deterministic equivalent reformulation along with a tailored Frank–Wolfe algorithm to efficiently solve our problem as well as a robust counterpart based on min-max regret. We present simulation-based numerical experiments from previous National Basketball Association seasons 2004–2019, and we show that our models are computationally efficient, outperform a greedy benchmark that approximates a nonrankings-based scheduling policy, and produce interpretable results.
Managerial implications: Our data-driven decision-making framework may be used to produce a shortened season with 25%–50% fewer games while still producing an end-of-season ranking similar to that of the full season, had it been played.
问题定义：职业体育联赛可能因各种原因暂停，例如最近的2019冠状病毒病大流行。当重新开放时，联赛必须解决的一个关键问题是如何在缩短的时间框架内适当选择剩余比赛的一部分来结束赛季。尽管有关从头开始安排整个赛季的文献丰富，但结束现有赛季则完全不同。我们的方法试图实现与完整赛季进行时可能产生的结果相似的球队排名。
方法/结果：我们提出了一个数据驱动模型，利用预测性和规定性分析来为赛季剩余部分制定一个由原定比赛的子集组成的赛程。我们的模型在随机优化模型中引入了基于新颖排名的目标，其参数首先使用预测模型进行估计。我们引入了一个确定性等价重构以及一个定制的Frank-Wolfe算法来有效解决我们的问题，以及一个基于最小最大遗憾的稳健对应物。我们展示了基于2004-2019年以前美国国家篮球协会赛季的模拟数值实验，并表明我们的模型在计算上是高效的，优于一个贪婪基准，该基准近似于一个非基于排名的赛程安排政策，并产生可解释的结果。
管理意义：我们的数据驱动决策框架可以用来在减少25%-50%比赛的同时，产生一个与完整赛季结束时相似的赛季末排名，如果它被进行了的话。
