Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WB31

6 - Impact of Facility Location and Inventory Allocation in a Blood Supply Chain Srimathy Mohan, Arizona State University, Department of Supply Chain Management, W.P. Carey School Of Management, Tempe, AZ, 85287-4706, United States, Mohan Gopalakrishnan The blood supply chain must be well coordinated and high performing. Using American Red Cross (ARC) data, this research focuses on analyzing the strategic decisions that are made when planning for non-blood inventories including, facility location and resource allocation decisions. We first formulate a fixed- charge model to minimize the total cost locating the non-blood supplies and transporting them to the demand points when required. We will also conduct sensitivity analyses for demand input and the fixed facility cost and analyze the impact on location and size decisions. We thus contribute to “management engineering” in the humanitarian logistics domain. Ride Sharing and Microtransit Sponsored: TSL/Urban Transportation Sponsored Session Chair: Changhyun Kwon, University of South Florida, Tampa, FL, 33620, United States 1 - Locality Sensitive Hashing Based Matching for Ride-Sharing Chinmoy Dutta, Lyft, Mountain View, CA, 94040, United States A standard approach to ride-matching is to generate all feasible matchings of rides, filter and score them, and solve a constraint optimization problem. For large ride-sharing platforms, generating all the feasible matchings is an impractical combinatorial problem. We propose a novel method to circumvent this combinatorial challenge using locality-sensitive hashing which lets us efficiently construct a high-quality set of feasible matches. 2 - A Practical Iterative Combinatorial Auction Design for Fractional Ownership of Autonomous Vehicles Aigerim Bogyrbayeva, University of South Florida, Tampa, FL, United States, Mahdi Takalloo, Hadi Charkhgard, Changhyun Kwon This study explores a market design for fractional ownership of autonomous vehicles (AVs), where an AV is co-owned by a group of individuals. In particular, a practical combinatorial auction design based on the well-known Combinatorial Clock Auction is proposed. The study discusses the unique features and challenges of the proposed auction and provides insights on the design performance based on numerical experiments. 3 - Computationally Efficient Truthful Mechanisms for Large-scale P2P Ridesharing Systems Neda Masoud, University of Michigan Ann Arbor, 2350 Hayward St., 2124 Gg Brown Bldg., Civil And Environmental Engineering, Ann Arbor, MI, 48109, United States, Roger Lloret We present a truthful Maximal-In-Distributional-Range (MIDR) randomized mechanism for NP-hard large-scale P2P ride-sharing problems, based on the Lavi- Swamy scaling decomposition technique. Instead of using the ellipsoid algorithm, we design a column generation scheme which makes the technique tractable for large-scale problems. Two pricing problems are tested: exponential exact and quasi-polynomial state-space relaxation. n WB30 North Bldg 221C Advances in Facility Location Algorithms Sponsored: TSL/Facility Logistics Sponsored Session Chair: Xin Wang, University of Wisconsin-Madison, Madison, WI, 53562-4278, United States 1 - Linear-Time Sweeping Algorithm for Discretization of Continuum Approximation Xiaopeng Li, University of South Florida, Tampa, FL, 33620, United States, Hongqiang Fan, Lifen Yun This paper proposes a construct-heuristic algorithm, the direct sweeping algorithm, to discretize a continuous location solution for the continuum approximation model. This algorithm is based on the idea of spatial searching and has only a linear computational complexity. Through the numerical examples, we study the performance of our proposed algorithm. The results of case studies indicate that this algorithm can solve the discrete facility locations problem more efficiently and has a robust performance compared with existing discretization algorithm. n WB29 North Bldg 221B

2 - Charging Electric Vehicle Sharing Fleet Long He, National University of Singapore, Mochtar Riady Building, BIZ1 8-73, 15 Kent Ridge Drive, Singapore, 119245, Singapore, Guangrui Ma, Wei Qi, Xin Wang In this paper, we develop models to jointly optimize the sites and sizes of charging stations, along with the coupled fleet charging and repositioning operations. We closely track EV energy levels and explicitly depict stochastic charging operations. We formulate the problem as a nonlinear optimization program and develop approximation optimization models as lower and upper bounds, which is computationally efficient. With real data sets of car2go’s operations in San Diego, we further conduct retrospective and futuristic case studies to discuss several managerial implications. 3 - Integrating Car-sharing with Public Transit Xin Wang, University of Wisconsin-Madison, Madison, WI, 53562- 4278, United States, Yikang Hua Car sharing service complements public transit with flexibility while reducing personal vehicles ownership. We build a continuous approximation model to help design car sharing infrastructure based on existing public transit network. We consider both the profitability of car sharing service and improvement of travelers’ utility. 4 - Performance Evaluation and Optimization of Reliable Sensor Deployment Zhoutong Jiang, University of Illinois, Champaign, IL, 61820, United States, Siyang Xie, Yanfeng Ouyang The accuracy of object positioning and surveillance depends on how multiple sensors are deployed in the sensor system and might suffer from probabilistic disruptions. To evaluate the performance of such system and optimize the sensor deployment strategy under different design criteria, we first develop a continuum approximation model to estimate the system cost and then optimize the sensor deployment strategies for different objectives. A series of numerical experiments are conducted to illustrate how the proposed model can be applied to provide near-optimum results within a very short computation time. n WB31 North Bldg 222A Transportation-Planning II Contributed Session Chair: Zhongxiang Wang, University of Maryland, Greenbelt, MD, 20770, United States 1 - A Hub-based Rebalancing Method for Free-floating Bike Sharing Systems Vahid Mahmoodian, PhD Student, University of South Florida, ENG 226, 4202 E. Fowler Ave, Tampa, FL, 33620, United States, Yu Zhang, Hadi Charkhgard In this study, a hub-based simulation model is presented to rebalance bikes in a free-floating bike sharing system. Such systems are one of the new transportation modes that received attention in recent years. This is highlighted by the fact that many companies with station based bike sharing systems are switching to free- floating systems. The proposed model takes advantage of both user and operator based rebalancing in order to maximize the service level and minimize the cost. The approach is applied to a real-world bike sharing setting and computational results show its efficacy. 2 - Taxonomy and Abstraction of the On-demand Mobility Aviation System for Smarter Cities Wancheng Dai, Florida Institute of Technology, 150 W. University Blvd, Melbourne, FL, 32901, United States, Rodrigo Mesa Arango On-demand mobility (ODM) is a new transportation system providing point-to- point air services to metropolitan travelers. ODM can significantly decrease travel times and potentially improve several ground-transportation externalities, like emissions and congestion. However, the complexity and novelty of future ODM systems impede its proper understanding for informed decisions. This work employs a system of systems framework to characterize the ODM system, and guide modeling efforts that can answer related questions by stakeholders of the system. 3 - A Cell-based Transportation Planning Model of Humanitarian Logistics and its Application Kazuaki Okubo, Associate Professor, Ehime University, 3 Bunkyo machi, Matsuyama city, Matsuyama, 790-8577, Japan, Pang-jo Chun We proposed an optimal transportation model for emergency relief supplies based on cell-based Merchant Nemhauser model. The model minimizes the total transportation time from supply origins to shelters and provides an optimal allocation of the supplies and labors in a dynamic setting. We applied the model to an earthquake disaster in Ehime Prefecture to evaluate the effect of changes in depot location. We show an additional depot enables the transportation in the shorter time. And it also improves the robustness of transportation against variation in the amount of supply and demand.

451