Academic Papers

Author(s)

Vincent W. Slaugh, Ludwig Dierks, Alan Scheller-Wolf, Andrew C. Trapp, M Utku Ünver

Summary

Federal, state, local, and nonprofit agencies in the United States spend more than $40 billion annually to serve vulnerable children. Interactions with the child welfare system can change children’s lives significantly, ideally protecting children from harm and helping families thrive. In a chapter in the edited volume Nonprofit Operations and Supply Chain Management: Theory and Practice, Tepper School of Business researcher Alan Scheller-Wolf and his co-authors suggest that improving operational decisions can be a critical factor in increasing the efficiency, effectiveness, and fairness of child welfare systems. The chapter describes challenges for managing processes at three salient stages of the child welfare system: investigations and pre-removal services, foster care, and adoption. The chapter highlights relevant operations-focused insights from child welfare researchers and the limited research on child welfare by economists and operations researchers. It also identifies important issues to consider in future research and connects them to studies on operations management from related domains. The chapter emphasizes that making better operational decisions can direct limited resources to families needing services, prevent unfounded investigations of blameless families, and help frontline child welfare workers in their critical and demanding jobs.

Author(s)

Yanhan (Savannah) Tang, Andrew Li, Alan Scheller-Wolf

Summary

Many consumers learn about new products through exposure and user experience, commonly referred to as consumer learning. Repeated exposure and increased knowledge of a product or ad can lead to increased liking, preference, or positive attitudes, known as the mere exposure effect. Marketers can leverage consumer learning and the mere exposure effect by running consistently personalized advertisements to encourage consumers to make purchases and build customer loyalty. In a new study, researchers formulated a multi-armed bandit (MAB) model and applied it to the allocation of split liver transplantation (SLT) and marketing at organ transplant centers. The model included provisions ensuring that the choices of arms were subject to fairness constraints to guarantee equity in liver transplantation or preserve variety in personalized marketing. The study’s algorithms had superior numerical performance compared to standard MAB algorithms settings in which learning through experience/the mere exposure effect and fairness/variety-seeking concerns exist. The use of this model has broad implications for SLT (e.g., using algorithms to help evaluate strategies to increase the proliferation of SLT) as well as for other applications (e.g., using algorithms to harness customer learning and the mere exposure effect to boost sales and brand loyalty).

Author(s)

Lin An, Andrew A. Li, Benjamin Moseley, R. Ravi

Summary

In a new study, researchers examined an updated version of the classic newsvendor problem. In the old model, a newsvendor has to select a quantity of inventory before seeing what the demand is for the inventory, with the demand randomly drawn from a known distribution. Motivated by applications (e.g., cloud provisioning, staffing), researchers considered a new model, which they termed the nonstationary newsvendor model, in which newsvendor-type decisions must be made sequentially in the face of demand drawn from a randomly determined process that is both unknown and nonstationary. The study analyzed the nonstationary newsvendor with and without predictions. Researchers validated their findings with experiments based on three actual data sets, proving that they closed nearly three-quarters of the gap between the best approaches based on nonstationarity and predictions alone.