论文导读 | Manufacturing & Service Operations Management近期文章精选

编者按

在本系列文章中，我们梳理了顶刊Manufacturing & Service Operations Management5月份发布有关OR/OM以及相关应用的文章之基本信息，旨在帮助读者快速洞察行业/学界最新动态。

推荐文章1

● 题目：Robust Drone Delivery with Weather Information

考虑天气信息的鲁棒无人机送货

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2022.0339

● 作者：Chun Cheng , Yossiri Adulyasak , Louis-Martin Rousseau

● 发布时间：

2024.05.03

● 摘要：

Problem definition: Drone delivery has recently garnered significant attention due to its potential for faster delivery at a lower cost than other delivery options. When scheduling drones from a depot for delivery to various destinations, the dispatcher must take into account the uncertain wind conditions, which affect the delivery times of drones to their destinations, leading to late deliveries.

Methodology/results: To mitigate the risk of delivery delays caused by wind uncertainty, we propose a two-period drone scheduling model to robustly optimize the delivery schedule. In this framework, the scheduling decisions are made in the morning, with the provision for different delivery schedules in the afternoon that adapt to updated weather information available by midday. Our approach minimizes the essential riskiness index, which can simultaneously account for the probability of tardy delivery and the magnitude of lateness. Using wind observation data, we characterize the uncertain flight times via a cluster-wise ambiguity set, which has the benefit of tractability while avoiding overfitting the empirical distribution. A branch-and-cut (B&C) algorithm is developed for this adaptive distributionally framework to improve its scalability. Our adaptive distributionally robust model can effectively reduce lateness in out-of-sample tests compared with other classical models. The proposed B&C algorithm can solve instances to optimality within a shorter time frame than a general modeling toolbox.

Managerial implications: Decision makers can use the adaptive robust model together with the cluster-wise ambiguity set to effectively reduce service lateness at customers for drone delivery systems.

问题定义：无人机送货由于其相较于其他送货选项具有更快且成本更低的潜力，近年来受到广泛关注。在从配送中心调度无人机进行各地送货时，调度员必须考虑不确定的风力条件，这会影响无人机到达目的地的送货时间，从而导致延误送货。

方法论/结论：为了减少由风力不确定性引起的送货延误风险，我们提出了一个两阶段的无人机调度模型，从而鲁棒性地优化送货时间表。在这个框架中，早晨做出调度决策，并预留不同的下午送货时间表，以适应中午更新的天气信息。我们的方法最小化了一个重要的风险指数，该指数可以同时考虑送货迟延的概率和迟延的程度。利用风力观测数据，我们通过集群模糊集来描述不确定的飞行时间，这样可以在避免过度拟合经验分布的同时保持可处理性。我们开发了一个Branch & Cut算法来提高这个自适应分布鲁棒框架的可扩展性。在样本外测试中，我们的自适应分布鲁棒模型相比其他传统模型能有效减少延迟。我们提出的B&C算法可以在比一般建模工具更短的时间内求解到最优解。

管理启示：决策者可以使用自适应鲁棒模型和集群模糊集，有效减少无人机送货系统在顾客服务上的延迟。

推荐文章2

● 题目：Frontiers in Operations: Valuing Nursing Productivity in Emergency Departments

运筹学前沿：评估急诊科护理生产力的价值

● 发布时间：2024.05.08

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2023.0039

● 作者：Hao Ding , Sokol Tushe , Diwas Singh KC , Donald K. K. Lee

● 摘要：

Problem definition: We quantify the increase in productivity in emergency departments (EDs) from increasing nurse staff. We then estimate the associated revenue gains for the hospital and the associated welfare gains for society. The United States is over a decade into the worst nursing shortage crisis in history fueled by chronic underinvestment. To demonstrate to hospital managers and policymakers the benefits of investing in nursing, we clarify the positive downstream effects of doing so in the ED setting.

Methodology/results: We use a high-resolution data set of patient visits to the ED of a major U.S. academic hospital. Time-dependent hazard estimation methods (nonparametric and parametric) are used to study how the real-time service speed of a patient varies with the state of the ED, including the time-varying workloads of the assigned nurse. A counterfactual simulation is used to estimate the gains from increasing nursing staff in the ED. We find that lightening a nurse’s workload by one patient is associated with a 14% service speedup for every patient under the nurse’s care. Simulation studies suggest that adding one more nurse to the busiest 12-hour shift of each day can shorten stays and avert $160,000 in lost patient wages per 10,000 visits. The reduction in service times also frees up capacity for treating more patients and generates $470,000 in additional net revenues for the hospital per 10,000 visits. Extensive sensitivity analyses suggest that our key message—that investing in nursing will more than pay for itself — is likely to hold across a wide range of EDs.

Managerial implications: In determining whether to invest in more nursing resources, hospital managers need to look beyond whether payer reimbursements alone are sufficient to cover the up-front costs to also account for the resulting downstream benefits.

问题定义：我们量化了通过增加护士人员数量在急诊科（ED）中提高生产力的效果。然后，我们估算了医院的相关收入增益以及社会的相关福利增益。美国已经进入了历史上护士短缺危机最严重的十年之久，这是由长期的投资不足所引发的。为了向医院管理者和决策者展示投资于护理的好处，我们阐明了在急诊科设置中这样做的积极后果。

方法论/结果：我们使用一家美国一流学术医院急诊科患者就诊的高精度数据集。我们采用时间相关的风险估计方法（非参数和参数方法）来研究患者实时服务速度与急诊科状态之间的变化，包括分配给护士的工作负载的时间变化。我们使用反事实模拟来估算增加急诊科护士人员的收益。我们发现，为一名护士的工作负荷减轻一个患者，可使护士照顾的每个患者的服务速度提高14%。模拟研究表明，在每天最繁忙的12小时班次中增加一名护士可以缩短住院时间，并节省每10,000次就诊中16万美元的患者误工费。服务时间的减少还可以为治疗更多患者释放出容量，并为医院每10,000次就诊产生47万美元的额外净收入。广泛的灵敏度分析表明，我们的主要观点——投资于护理将更加划算——可能适用于各种急诊科。

管理启示：在决定是否投资更多护理资源时，医院管理者除了考虑支付方报销是否足以覆盖前期成本以外，还需要考虑到由此产生的下游收益。

推荐文章3

● 题目：The Effects of Selling Formats and Upstream Competition on Product Pricing and Quality Design

销售形式和上游竞争对产品定价和质量设计的影响

● 发布时间：2024.05.14

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2022.0470

● 作者：Lu Hsiao , Xin Ma , Ying-Ju Chen

● 摘要：

Problem definition: In practice, consumers are directly or indirectly connected with branded manufacturers through intermediaries across industries. In this paper, we explore the effects of different selling formats on product quality and price depending on consumer valuations in a market. We employ a distribution family to comprehensively capture the heterogeneity of consumer valuations. Motivated by realistic phenomena, consumer valuations are used to investigate strategic decisions under different selling formats that are not trivial to analyze.

Methodology/results: We develop game-theoretical models to examine the equilibrium decisions of stakeholders. The impact of consumer valuations is investigated and validated using sensitivity analysis, and the results are connected to practice. First, we find that agency selling induces a premium quality and maximizes the channel profit; remarkably, a nonmonotonic (approximate U-shaped) relationship exists between the agency fee and consumer valuations. A higher consumer surplus can be achieved in an agency selling channel compared with a reselling channel, particularly when targeting a mass of high-end consumers. Second, by examining distinct consumer valuations, maintaining top-notch quality and the highest price in an agency selling channel is not universally viable under some conditions. Third, in the case of production-level competition, an agency selling format tends to cause product quality to vary noticeably. Moreover, in the hybrid selling channel, in contrast to agency selling, the high-type manufacturer reduces both quality and price, which bolsters the overall profits of the channel and the consumer surplus.

Managerial implications: Branded manufacturers can efficiently respond to individualized consumer needs in a centralized distribution channel. In contrast, for selling basic products, the reselling channel could contribute to achieving economies of scale and offering competitive prices. In the agency selling channel, standardized pricing determined by branded manufacturers can create a consistent perception of product quality throughout the distribution network.

问题定义：在实践中，消费者通过跨行业的中介直接或间接地与品牌制造商相连。本文探讨了在市场中根据消费者估值探讨不同销售形式对产品质量和价格的影响。我们采用一个分布族来全面捕捉消费者估值的异质性。受现实现象的启发，消费者估值被用来研究在不同销售形式下的战略决策，只是这些决策并不容易分析。

方法论/结果：我们开发了博弈论模型来研究利益相关者的均衡决策。通过灵敏度分析来研究和验证消费者估值的影响，并将结果与实践联系起来。首先，我们发现代理销售导致了高品质，并最大化了渠道利润；值得注意的是，在代理费用和消费者估值之间存在非单调（近似U形）关系。与再销售渠道相比，面向大量高端消费者时在代理销售渠道中可以实现更高的消费者剩余。其次，通过研究不同的消费者估值，我们发现保持顶级品质和最高价格在代理销售渠道中并非普遍可行。第三，在生产水平竞争的情况中，代理销售形式往往会导致产品质量明显变化。此外，在混合销售渠道中，与代理销售相比，高端制造商降低了品质和价格，从而增加了整个渠道和消费者剩余的总体利润。

管理启示：品牌制造商可以在集中式分销渠道中有效地满足个性化消费者需求。相比之下，对于销售基本产品，再销售渠道有助于实现规模经济并提供具有竞争力的价格。在代理销售渠道中，由品牌制造商确定的标准定价可以在整个分销网络中创建一致的产品质量感知。

推荐文章4

● 题目：Frontiers in Operations: Employees vs. Contractors: An Operational Perspective

运筹前沿：员工与承包商：一个运营视角

● 发布时间：2024.05.17

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2023.0029

● 作者：Ilan Lobel , Sébastien Martin , Haotian Song

● 摘要：

Problem definition: We consider a platform’s problem of how to staff its operations given the possibilities of hiring employees and setting up a contractor marketplace. We aim to understand the operational difference between these two work arrangement models.

Methodology/results: We consider a model where demand is not only stochastic but also evolving over time, which we capture via a state of the world that determines the demand distribution. In the case of employees, the platform controls the number of employee hours it uses for serving demand, whereas in the case of contractors, it sets the wage paid to them per utilized hour. We show that although the employee problem is equivalent to a standard newsvendor, the contractor one corresponds to an unusual version of the newsvendor model where utilization is the control variable.

Managerial implications: This distinction makes the contractor model more flexible, allowing us to prove that it performs significantly better, especially if the order of magnitude of demand is unknown. Meanwhile, hybrid solutions that combine both employees and contractors have complex optimal solutions and offer relatively limited benefits relative to a contractor marketplace.

问题定义：考虑到雇佣员工和建立承包商市场的可能性，我们考虑了一个平台如何组织其运营工作的问题。我们旨在理解这两种工作安排模式之间的运营差异。

方法论/结果：我们考虑了一个模型，其中需求不仅是随机的，而且随着时间的推移而发展进化，我们通过决定需求分布的普遍状态来捕捉这一点。在雇佣员工的情况中，平台控制其用于满足需求的员工工时数量，而在承包商的情况中，它设定给予他们的每个利用小时的工资。我们表明，员工问题等价于标准的报童模型，而承包商问题等价于一个不常规的报童版本，其中利用率是控制变量。

管理启示：这一区分使承包商模型更加灵活，我们证明了它在需求的数量级是未知的情况下表现得明显更好，与此同时，将员工和承包商结合的混合解决方案具有复杂的最优解，然而相对于承包商市场，其提供的好处相对有限。

推荐文章5

● 题目：Pooling and Boosting for Demand Prediction in Retail: A Transfer Learning Approach

零售需求预测中的汇聚和增强：一种迁移学习方法

● 发布时间：2024.05.30

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2022.0453

● 作者：Dazhou Lei , Yongzhi Qi, Sheng Liu , Dongyang Geng, Jianshen Zhang, Hao Hu, Zuo-Jun Max Shen

● 摘要：

Problem definition: How should retailers leverage aggregate (category) sales information for individual product demand prediction? Motivated by inventory risk pooling, we develop a new prediction framework that integrates category-product sales information to exploit the benefit of pooling.

Methodology/results: We propose to combine data from different aggregation levels in a transfer learning framework. Our approach treats the top-level sales information as a regularization for fitting the bottom-level prediction model. We characterize the error performance of our model in linear cases and demonstrate the benefit of pooling. Moreover, our approach exploits a natural connection to regularized gradient boosting trees that enable a scalable implementation for large-scale applications. Based on an internal study with JD.com on more than 6,000 weekly observations between 2020 and 2021, we evaluate the out-of-sample forecasting performance of our approach against state-of-the-art benchmarks. The result shows that our approach delivers superior forecasting performance consistently with more than 9% improvement over the benchmark method of JD.com. We further validate its generalizability on a Walmart retail data set and through alternative pooling and prediction methods.

Managerial implications: Using aggregate sales information directly may not help with product demand prediction. Our result highlights the value of transfer learning to demand prediction in retail with both theoretical and empirical support. Based on a conservative estimate of JD.com, the improved forecasts can reduce the operating cost by 0.01–0.29 renminbi (RMB) per sold unit on the retail platform, which implies significant cost savings for the low-margin e-retail business.

问题定义：零售商如何利用汇总（类别）销售信息进行个别产品需求预测？受库存风险汇集的启发，我们开发了一个新的预测框架，该框架整合了类别-产品销售信息，以利用Pooling汇总的优势。

方法论/结果：我们提出在一个迁移学习框架中结合不同聚合级别的数据。我们的方法将顶层销售信息视为底层预测模型的正则化项。我们表征了我们模型在线性情况下的误差性能，并展示了Pooling的好处。此外，我们的方法利用了与正则化梯度增强树的自然连接，从而实现了大规模应用的可扩展实现。基于与京东的内部研究，涵盖了2020年至2021年间超过6,000个周观察数据，我们评估了我们方法在样本外的预测性能，并与最先进的基线进行了对比。结果显示，我们的方法在京东的基线方法上持续提供了优越的预测性能，提高了超过9%。我们进一步在沃尔玛零售数据集上验证了其泛化性能，以及通过其他汇集和预测方法。

管理启示：直接使用汇总销售信息可能不利于产品需求预测。我们的结果突显了迁移学习在零售需求预测中的价值，具有理论和经验支持。根据京东的保守估计，改进的预测可以将零售平台上每单位的运营成本降低0.01至0.29人民币（RMB），这意味着对于低利润的电子零售业务而言，可实现显著的成本节省。

推荐文章6

● 题目：Should Only Popular Products Be Stocked? Warehouse Assortment Selection for E-Commerce Companies

应该只储存热门产品吗？电子商务公司的仓库商品选择

● 发布时间：2024.05.30

● 原文链接：https://pubsonline.informs.org/doi/abs/10.1287/msom.2022.0428

● 作者：Xiaobo Li , Hongyuan Lin , Fang Liu

● 摘要：

Problem definition: This paper studies the single-warehouse assortment selection problem that aims to minimize the order fulfillment cost under the cardinality constraint. We propose two fulfillment-related cost functions corresponding to spillover fulfillment and order splitting. This problem includes the fill rate maximization problem as a special case. We show that although the objective function is submodular for a broad class of cost functions, the fill rate maximization problem with the largest order size being two is NP-hard.

Methodology/results: To make the problem tractable to solve, we formulate the general warehouse assortment problem under the two types of cost functions as mixed integer linear programs (MILPs). We also provide a dynamic programming algorithm to solve the problem in polynomial time if orders are nonoverlapping. Furthermore, we propose a simple heuristic called the marginal choice indexing (MCI) policy that allows the warehouse to store the most popular products. This policy is easy to compute, and hence, it is scalable to large-size problems. Although the performance of MCI can be arbitrarily bad in some extreme scenarios, we find a general condition under which it is optimal. This condition is satisfied by many multi-purchase choice models.

Managerial implications: Through extensive numerical experiments on a real-world data set from RiRiShun Logistics, we find that the MCI policy is surprisingly near optimal in all the settings we tested. Simply applying the MCI policy, the fill rate is estimated to improve by 9.18% on average compared with the current practice for the local transfer centers on the training data set. More surprisingly, the MCI policy outperforms the MILP optimal solution in 14 of 25 cases on the test data set, illustrating its robustness against demand fluctuations.

问题定义：本文研究了在基数约束下最小化订单履行成本的单一仓库组合选择问题。我们提出了两种与履行相关的成本函数，分别对应于溢出履行和订单拆分。该问题将填充率最大化问题囊括进来作为一种特殊情况。我们证明了尽管目标函数对于广泛类别的成本函数是子模的，但最大订单大小为两个的填充率最大化问题是 NP-Hard的。

方法/结果：为了使问题易于解决，我们将在两种类型的成本函数下的一般仓库组合问题表述为混合整数线性规划（MILP）。如果订单不重叠，我们还提供了一个动态规划算法来在多项式时间内解决该问题。此外，我们提出了一种简单的启发式算法，称为边际选择索引（MCI）策略，允许仓库存储最受欢迎的产品。这个策略易于计算，因此对于大规模问题是可扩展的。尽管在一些极端情况下，MCI 的性能可能会极差，但我们发现了一个一般条件，在该条件下它是最优的，此外这个条件被不少多重选购选择模型所满足。

管理意义：通过对来自日日顺物流的真实数据集进行大量数值实验，我们发现 MCI 策略在我们测试的所有设置中都令人惊讶地接近最优。仅应用 MCI 策略，据估计，与当前实践相比，训练数据集中的地方转运中心的填充率平均提高了 9.18%。更令人惊讶的是，MCI 策略在测试数据集中的 25 个案例中有 14 个超过了 MILP 最优解，说明它对需求波动具有很强的鲁棒性。