INFORMS 2021 Program Book

INFORMS Anaheim 2021

WB32

implications ambiguous. We propose a way to refine the current AB5. 2 - Inventory and Pricing Optimization for Resale Firms Emily C. Barbee, University of Alabama, Tuscaloosa, AL, United States, BurcuB. Keskin Recent growth in e-commerce and sustainability has fueled demand for resale. Resale firms source used goods from consumers online. Supply is uncertain and item quality varies. We model this unique context as a joint inventory and pricing problem. We investigate responsive and committed pricing under price- and quality-dependent demand. WB34 CC Room 209B In Person: Transportation-Freight and Logistics Contributed Session Chair: Giulia Burchi, DecisionBrain, Paris, 75010, France 1 - The End of “Set It and Forget It’’ Pricing? Opportunities For Market-based Freight Contracts Angela Acocella, Massachusetts Institute of Technology, Cambridge, MA, United States, Chris Caplice, Yossi Sheffi In the for-hire truckload market, firms experience unexpected costs from contracted transportation service providers due to load rejections. Moreover, the dominant procurement strategy results in long-term fixed-price contracts that become stale as providers’ networks change and freight markets fluctuate between over and under supply. We build behavioral models of carriers’ load acceptance decisions under two distinct market conditions. We quantify carriers’ contract price stickiness as their best-known alternative priced load options become more attractive for different lane, freight, and carrier segments to identify best opportunities for market-based contracts. 2 - A Holistic Approach for Intermodal Facility Location and Freight Distribution under Hurricane Disruptions Vishal Badyal, Clemson University, Clemson, SC, United States, William G. Ferrell, Nathan Huynh, Bhavya Padmanabhan We study the intermodal facility location problem under hurricane disruptions. Hurricanes can cause disruption in supply at shippers and throughput capacity at intermodal facilities. Realistic hurricane scenarios are generated using k-means clustering. A level method-based decomposition solution approach is applied. The model is tested and validated by developing a case study for the state of South Carolina. Real-world data sets (FAF4 and HURDAT2) are used. The results show that as direct shipping costs increase, the long-term savings using this model increase non-linearly. The increase in direct shipping cost leads to more intermodal locations selected despite being partially disrupted. 3 - Inbound Logistics Optimization Solution for Toyota Giulia Burchi, DecisionBrain, Paris, France In this presentation, DecisionBrain will talk about an inbound logistics optimization solution for Toyota, which resulted in over 10% cost reductions. The project was completed in 8 months, from conception to go-live. This produced a high ROI and a payback time of less than one year. The solution focused on optimizing the Orders Grouping, Trucks Routing: and 3D packing.The solution was built on top of IBM Cplex Optimization Studio and on IBM Decision Optimization Center (DOC), which allowed for a fast and effective implementation, from design to deployment, delivering significant ROI. WB35 CC Room 210A In Person: Tackling Emerging Logistics Challenges with Large-scale Analytics General Session Chair: Alexandria Schmid, MIT, Somerville, MA, 02143, United States 1 - Submodular Dispatching Ignacio Erazo, ISyE Georgia Tech, Atlanta, GA, United States, Alejandro Toriello We introduce a submodular dispatching model motivated by applications in e- commerce distribution and scheduling, among others. A server must process a set of jobs to minimize the makespan; jobs have release times and the server is dispatched to process jobs in batches, where the batch dispatching time is non- decreasing and submodular. We prove that the general problem is strongly NP-hard, and characterize “FIFO-optimal” processing time functions for which an efficient dynamic program is optimal. The algorithm produces the optimal batch selection in which jobs are processed in FIFO order, which also serves as a heuristic for the general case, where we show that it has a 1.5 approximation ratio. We also study the lower bound provided by a column generation LP, and verify the efficacy of our heuristic and bound in computational experiments.

WB32 CC Room 208B In Person: New Directions in Pricing and Auction Design General Session Chair: Amine Allouah, Columbia University, New York, NY, 10027, United States 1 - Optimal Auction Design with Deferred Inspection and Reward Azarakhsh Malekian, University of Toronto, Toronto, ON, 02143- 2434, United States, Saeed Alaei, Alexandre Belloni, Ali Makhdoumi Consider a mechanism run by an auctioneer who can use both payment and inspection instruments to incentivize agents. The timeline of the events is as follows. Based on a pre-specified allocation rule and the reported values of agents, the auctioneer allocates the item and secures the reported values as deposits. The auctioneer then inspects the values of agents and, using a pre-specified reward rule, rewards the ones that have reported truthfully. Using techniques from convex analysis and calculus of variation, for any distribution of values, we fully characterize the optimal mechanism for a single agent. Using Border’s theorem and duality, we find conditions under which our characterization extends to multiple agents. The optimal allocation function is not a thresholding strategy and instead is an increasing and continuous function of the types. 2 - Persuading Customers to Buy Early: The Value of Personalized Information Provisioning Ramandeep Randhawa, University of Southern California, Los Angeles, CA, 90089-1035, United States, Kimon Drakopoulos, Shobhit Jain We study a pricing and information provisioning game between a better informed seller (such as a retailer) and its customers. The seller is (ex-post) better informed about product availability and can choose how to communicate this information to the customers. Using a Bayesian persuasion framework, we find that public information provisioning in which the firm sends the same information to all customers has limited value. However, personalized information provisioning, in which the firm can share different information with different customers, has significant value and has attributes very similar to personalized pricing. 3 - Optimal Pricing with a Single Point Achraf Bahamou, Columbia University, Columbia, NY, United States, Amine Allouah, Omar Besbes We study the following fundamental data-driven pricing problem. How can/should a decision-maker price its product based on observations at a single historical price? The decision-maker optimizes over (potentially randomized) pricing policies to maximize the worst-case ratio of the revenue she can garner compared to an oracle with full knowledge of the distribution of values when the latter is only assumed to belong to a broad non-parametric set. In particular, our framework applies to the widely used regular and monotone non-decreasing hazard rate (mhr) classes of distributions. For settings where the seller knows the exact probability of sale associated with one historical price or only a confidence interval for it, we fully characterize optimal performance and near-optimal pricing algorithms that adjust to the information at hand. The framework we develop is general and allows to characterize optimal performance for deterministic or more general randomized mechanisms, and leads to fundamental novel insights on the value of information for pricing. As examples, against mhr distributions, we show that it is possible to guarantee 85% of oracle performance if one knows that half of the customers have bought at the historical price, and if only 1% of the customers bought, it is still possible to guarantee 51% of oracle performance. WB33 CC Room 209A In Person: Joint Inventory and Pricing Models General Session Chair: Emily Barbee, University of Alabama, Tuscaloosa, AL, United States 1 - Implications of Worker Classification in On-Demand Economy Zhoupeng (Jack) Zhang, Rotman School of Management, University of Toronto, Toronto, ON, M5 S. 3E6, Canada, Ming Hu, Jianfu Wang Workers in the gig economy have long been treated as independent contractors, which disqualifies them from employee benefits. We evaluate the impacts of California Assembly Bill 5 (AB5), a statute that requires on-demand platforms to reclassify their workers as employees. We model the service process of such a platform as a queueing system with long-term (LT) and ad hoc (AH) workers. We show that AB5 does not always improve LT workers’ welfare because, in the free market, the presence of AH workers can incentivize the company to pay a high piece-rate wage. While the company’s profit always decreases, transaction volume can either increase or decrease due to AB5, rendering consumer welfare

156