Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WB21

2 - Organizational Learning in Process Improvement Projects Venkat Venkateswaran, Georgia Institute of Technology, Room 4143, 800 West Peachtree, NW, Atlanta, GA, 30308, United States Venkat Venkateswaran, University of Illinois, Urbana-Champaign, IL, United States, Edward Arnheiter Many organizations engage in continuous process improvement activities. We study the learning that accrues when workers gain experience from such exercises. We use panel data from a large oil and natural gas extraction company to track the performance of 56 trained lean Six Sigma facilitators completing 233 projects over five years. The projects are all different, yet learning is discernible. Examples where trained personnel undertake activities not necessarily identical (the subject of traditional research) but falling within some discipline occur often in practice. For such applications, we propose a way to measure learning. The study data exhibits learning at an effective rate of about 85%. 3 - Optimal Decisions for Contract Farming under Weather Risk Xinping Wang, Nanjing Agricultural University, Nanjing, China, Shengnan Sun Agriculture has been greatly impacted by weather condition. In this paper, we consider a manufacturer who takes a farm crop as a raw material and transforms it into a finished product. The manufacturer encourages local farmers to grow the farm crop by offering them contracts to guarantee her supply under weather risk. Optimal decisions for two types of contracts, contract without reneging and contract with reneging, are analyzed. In addition, we numerically compare the optimal decisions of the company with different contracts and investigate the sensitivity of the results to the model parameters, such as weather condition. Design of Two-sided Marketplaces Sponsored: Revenue Management & Pricing Sponsored Session Chair: Siddhartha Banerjee, Cornell University, Ithaca, NY, 14853, United States Co-Chair: Pengyu Qian, Columbia University-Graduate School of Business, New York, NY, 10027, United States Co-Chair: Yashodhan Kanoria, Columbia Business School, New York, NY, United States 1 - Simple Mechanisms for Two-sided Markets Yang Cai, Yale University, New Haven, CT, United States We design simple mechanisms to approximate the Gains from Trade (GFT) in two-sided markets with multiple unit-supply sellers and multiple unit-demand buyers. A classical impossibility result by Myerson and Satterthwaite showed that even with only one seller and one buyer, no Individually Rational (IR), Bayesian Incentive Compatible (BIC) and Budget-Balanced (BB) mechanism can achieve full GFT (trade whenever buyer’s value is higher than the seller’s cost). On the other hand, they proposed the “second-best mechanism that maximizes the GFT subject to IR, BIC and BB constraints, which is unfortunately rather complex for even the single-seller single-buyer case. Our mechanism is simple, IR, BIC and BB, and achieves 1/2 of the optimal GFT among all IR, BIC and BB mechanisms. Our result holds for arbitrary distributions of the buyers’ and sellers’ values and can accommodate any downward-closed feasibility constraints over the allocations. 2 - Descending Price Optimally Coordinates Search Bo Waggoner, University of Pennsylvania, Philadelphia, PA, United States, Robert Kleinberg, Glen Weyl Contrary to common practice in selling homes and start-ups, mechanism design theory typically recommends English (increasing price) over Dutch (decreasing price) auctions. Yet this theory neglects the uncertain investment required to investigate purchases. We show that English and other standard auctions burden such investments with further uncertainty about the price necessary to win, potentially eliminating all gains from trade. In contrast, Dutch auctions preserve their properties without information costs because they guarantee, at the moment when investigation is optimal, a price at which the good can be purchased. 3 - State Dependent Control of Closed Queueing Networks Pengyu Qian, Columbia University-Graduate School of Business, New York, NY, 10027, United States, Siddhartha Banerjee, Yash Kanoria We study the state dependent control for a closed queueing network model inspired by ridesharing systems. The platform can choose which supply unit to assign to an incoming customer; the unit becomes available again at customer’s destination. The platform aims to minimize proportion of dropped demand. We propose a family of simple policies called Scaled MaxWeight (SMW) and prove that under complete resource pooling (analogous to Hall’s condition), any SMW policy induces an exponential decay of demand-drop probability as the number of supply units goes to infinity. We further show that there is an SMW policy that n WB19 North Bldg 128B

achieves the optimal exponent among all assignment policies and specify it analytically. 4 - Asymptotically Optimal Thickness of a Centralized Dynamic Matching Market with Independent and Identically Distributed Utilities Lawrence M. Wein, Stanford University, Graduate School of Business, 655 Knight Way, Stanford, CA, 94305-5015, United States, Martin I. Reiman We consider a centralized dynamic matching market, where buyers and sellers arrive according to independent Poisson processes with the same rate, and abandon (if not matched) after an exponential amount of time with the same mean. The matching utility between any buyer and seller is an iid random variable. Under a large-market asymptotic regime, we use extreme value theory and Markov chain asymptotics to show how the optimal market thickness varies according to the domain of attraction of the underlying matching utility distribution. n WB21 North Bldg 129B Revenue Management with Reusable Resources Sponsored: Revenue Management & Pricing Sponsored Session Chair: Cong Shi, University of Michigan, Ann Arbor, MI, 48105, United States 1 - Optimal Pricing Policy for Reusable Resource Systems with Strategic Customers Yiwei Chen, University of Cincinnati, Cincinnati, OH, United States, Cong Shi We consider a service system with a finite number of homogeneous reusable resources that dynamically serve customers. Customers stochastically arrive to the service system with heterogeneous requested service time intervals (allowing for advance reservations) and per unit of time service valuations. Customers are forward looking who strategize their purchase times. We propose a simple pricing policy that is fixed for each unit of time. This policy incentivizes customers to behave myopically. We show that this policy is asymptotically optimal. 2 - Real-time Dynamic Pricing for Revenue Management with Reusable Resources, Advance Reservation, and Deterministic Service Time Requirements Yanzhe Lei, Queen’s Univerisity, Kingston, ON, Canada, Stefanus Jasin We study a dynamic pricing problem where a firm uses a finite amount of resources to serve price-sensitive customers arriving randomly over time. Each customer may request to consume a combination of different types of resources for a deterministic duration of time, after which the resource can be immediately used to serve new customers. We develop real-time pricing controls and show that they are near-optimal in the regime of large demand and large resource capacity. We further extend our result to a more general setting with We consider the problem of dynamic pricing and assortment optimization for reusable resource in continuous time under time-varying demand. We develop a time-discretization strategy that yields a constant factor performance guarantee relative to the optimal policy. Our policies are computationally feasible and asymptotically optimal for large systems. Additionally, we develop heuristic methods that implement a bid-price strategy that accounts for the changing value of resources over time. 4 - Dynamic Assortment Optimization for Reusable Products with Random Usage Durations Mika Sumida, Cornell Tech, New York, NY, 10128, United States, Paat Rusmevichientong, Huseyin Topaloglu We consider dynamic assortment problems with reusable products, in which each customer chooses a product within an offered assortment, uses the product for a random duration of time, and returns the product back to the firm to be used by other customers. The goal is to find a policy for deciding on the assortment to offer to each customer so that the total expected revenue over a finite selling horizon is maximized. We present a policy that is guaranteed to obtain at least 50% of the optimal total expected revenue. The policy is based on constructing linear approximations to the optimal value functions and is computed through an efficient backward recursion. heterogeneous service time and advance reservation. 3 - Reusable Revenue Management at Scale Zachary Owen, David Simchi-Levi

447