Informs Annual Meeting 2017

MB04

INFORMS Houston – 2017

MB03C Grand Ballroom C MSOM Special Speaker: David Simchi-Levi Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Tunay Tunca, Unversity of Maryland, 8803 Courts Way, Silver Spring, MD, 20910, United States, ttunca@rhsmith.umd.edu 1 - Online Resource Allocation with Applications to Revenue Management David Simchi-Levi, Massachusetts Institute of Technology, Dept of Civil and Environmental Engineering, 77 Massachusetts Avenue Rm 1-171, Cambridge, MA, 02139, United States, dslevi@mit.edu Online resource allocation is a fundamental problem in OR and CS with applications such as offering products to customers, distributing jobs to candidates, assigning advertisers to ad slots, and matching drivers to passengers. These problems can be abstracted as follows: there are fixed resources, each of which can be sold at multiple known prices. These resources must be allocated on-the-fly, without assuming anything about future demand. In this talk we cover the CS and OR literature on the problem and in particular focus on two techniques: exploration and exploitation methods, as well as competitive analysis. Chair: Elena Belavina, University of Chicago Booth School of Business, Chicago, IL, 60637, United States, elena.belavina@chicagobooth.edu 1 - Design of Sales Force Compensation Schemes to Reduce Product Waste Arzum E. Akkas, Boston University, 100 Memorial Drive, 8-11C, Cambridge, MA, 02142, United States, aakkas@bu.edu We empirically examine sales-force compensation schemes of manufacturers that impact expiration of consumer-packaged-goods (CPG) at retailers to identify incentive-aligning compensation schemes. Using sales commission data from a CPG manufacturer, we investigate the impact of counterfactual schemes on manufacturer’s profit and product waste based on our model of sales-force decision process. 2 - Dynamic Staffing of Volunteer Gleaning Operations Erkut Sonmez, Boston University, 1496 Beacon St, Apt 4, Brookline, MA, 02446, United States, sonmez@bu.edu, Baris Ata, Deishin Lee Gleaning programs organize volunteer gleaners to harvest leftover crops that are donated by farmers for the purpose of feeding food-insecure individuals. Thus, the gleaning process simultaneously reduces food waste and food insecurity. However, the operationalization of this process is challenging because gleaning relies on two uncertain sources of input: the food and labor supplies. We develop a model to capture the uncertainties in food and labor supplies and seek a dynamic volunteer staffing policy that maximizes the long run average volume of food gleaned. 3 - Economically Motivated Adulteration in Farming Supply Chains: the Role of Supply Chain Dispersion and Traceability Somya Singhvi, MIT, 235 Albany Street, Cambridge, MA, 02139, United States, ssinghvi@mit.edu, Retsef Levi, Yanchong Zheng We develop a game-theoretic analytical model to study the effects of quality uncertainty, supply chain dispersion, traceability, and testing accuracy on economically motivated adulteration (EMA) in a farming supply chain with a distributed network of farmers. We further analyze the manufacturer’s optimal investment in traceability and testing frequency to satisfy a risk constraint while minimizing costs. We leverage the analysis of our model to derive important and new insights that can inform both policy makers and commercial entities in food supply chains to address EMA risks. We also use data from real world case studies to calibrate our model and highlight its ‘predictive’ power. MB04 320A Sustainable Food Supply and Food Waste Sponsored: Manufacturing & Service Oper Mgmt, Sustainable Operations Sponsored Session

3 - Mixing Time and Structural Inference for Bernoulli Auto-regressive Processes R. Srikant, UIUC, Urbana, IL, United States, rsrikant@illinois.edu, Carolyn Beck, Dimitris Katselis We introduce a multivariate random process called a Bernoulli Autoregressive Process (BAR). A BAR process models a discrete-time vector random process where the components of the vector are represented by the nodes in a graph. We show that the BAR process mixes rapidly, by proving that the mixing time is O(log p). For a network with p nodes, where each node has in-degree at most d<

165

Made with FlippingBook flipbook maker