Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

POSTER COMPETITION

67 - Controlling Freeway Merge Operations Under Conventional and Automated Vehicles Traffic Aschkan Omidvar, University of Florida, Gainesville, FL, 32603 United States, Mahmoud Pourmehrab, Lily Elefteriadou In this paper, we present an optimization algorithm for freeway operations at merge zones which maximizes the average speed of the segment in the presence of Automated Vehicles (AV) and human-operated (i.e., conventional) vehicles. The necessary algorithms are developed to simulate and carry out the merging operations on a 2-lane freeway (one merge line and one ramp line) which is tested under a variety of scenarios ranging in demand level, demand splits, and AV penetration rate. Results suggest that the proposed algorithm can efficiently manage the traffic at freeway merge zones and reduce average total travel time (increase average speed). 68 - Vulnerability Analysis and Assessment of Urban Rail Transit Network Meiyi Zhao, Southwest Jiaotong University, Chengdu, China, Kuniaki Sasaki, Guofang Li, Jianmei Zhu It is necessary to study vulnerability, what kind of impact the interruption will cause on urban rail transit network, described by the number of passengers multiply time loss, if emergency occurs. First, topological network is built by stations and lines. Then K paths are being searched by GA and their validity judged by time impedance threshold and train schedule. Next, passenger flow will be assigned to multiple paths. Normal distribution is chosen to describe the passenger behavior based on sensitivity to the shortest path and ratio will be corrected according to transfer times. After interruption, passengers were divided into 3 type. 1. all invalid; 2. partly invalid;3. all valid. Chair: Junming Yin, University of Arizona, Management Information Systems Department, McClelland Hall, Room 430BB, Tucson, AZ, 85721, United States Co-Chair: Neng Fan, University of Arizona, Tucson, AZ, 85721, United States Co-Chair: Burcu B. Keskin, University of Alabama, Tuscaloosa, AL, 35406, United States 1 - Smart Additive Manufacturing Using Data Analytics and Feedback Control Chenang Liu, Virginia Tech, Blacksburg, VA, 24061, United States, Zhenyu Kong A major challenge in additive manufacturing (AM) is to ensure product quality and consistency. To address this challenge, this research proposes an integrated manifold learning algorithm to achieve online quality monitoring and a bilateral time series modeling approach to implement online quality forecasting. With the application of feedback control system, the monitoring and forecasting results are also utilized to achieve significant quality improvement of AM via online process adjustment. The proposed methods demonstrate effective performance in real- world case studies. 2 - Distribution Network Reconfiguration with Decentralized Autonomous Electric Vehicles Zhaomiao Guo, Universtiy of Central Florida, Orlando, FL, 32816, United States Autonomous electric vehicles (AEVs) provide unique opportunities to cope with the uncertainties of distributed energy generations in distribution network. We investigate the potential benefits of dynamic distribution network reconfiguration (DDNR) considering AEVs’ spatial-temporal availability and their charging demand. A mixed integer programming model is proposed to optimally coordinate the charging/discharging of AEVs with DDNR, while satisfying AEVs’ original travel plan. Numerical studies show that DDNR and AEV are complemented to each other to improve distribution system operation. 3 - Fireline Construction in a Heterogeneous Forest Landscape Xu Yang, Northeastern University, Boston, MA, 02115, United States, Emanuel Melachrinoudis In this research we develop a methodology to construct the fireline in minimum time by considering the realistic case of heterogeneous forest landscape. We represent the forest landscape as a partition into homogeneous areas and consider multiple line segments crossing each polygon that represent potential paths of the fireline. This discretization reduces the forest landscape to a network whose nodes have as attribute the time of the fire arrival. A methodology is developed that uses a modified Dijkstra’s algorithm to find the fastest yet safe paths for two Tuesday, 12:30PM - 2:30PM n Tuesday Poster Competition North Exhibit Hall E Tuesday Poster Competition Competition Poster Session

firefighting crews who work simultaneously in two opposite directions until they meet and thus they encircle and contain the fire. 4 - Cell-based Network Flow Model Under Uncertainty in Evacuees’ Behavior by Social Influence Hyeong Suk Na, PhD Candidate, The Pennsylvania State University, University Park, PA, 16802, United States, Necdet Serhat Aybat, Soundar Kumara We investigate a stochastic network flow model for planning a large-scale hurricane evacuation strategy considering uncertainty on the number of vehicles leaving from the affected areas. We model human evacuation behavior using a time inhomogeneous discrete time Markov chain and this corresponds to using a stochastic S-shaped response curve. The proposed model is developed based on the Cell Transmission Model with joint chance constraints and it is handled by the sample average approximation method and Monte Carlo simulation techniques. A numerical case study examines the applicability of the proposed model and the effect of the evacuation traffic flows on different social network topologies. 5 - An Iterative Combinatorial Auction for Fractional Ownership of Autonomous Vehicles Aigerim Bogyrbayeva, University of South Florida, Tampa, FL, 33613, 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-leased by a group of individuals. We propose an iterative combinatorial auction design for this novel application. The study discusses the unique features of the proposed auction and delivers solution tools. 6 - Long Ties Accelerate Realistic Complex Contagions Amin Rahimian, Postdoctoral Associate, Massachusetts Institute ot Technology, Cambridge, MA, 02142, United States, Dean Eckles, Elchanan Mossel, Subhabrata Sen The spread of behaviors (i.e. social contagion) depends on the structure of the network of contacts between people. For simple contagion models borrowed from epidemic spread, highly clustered networks slow spread compared with more random networks, such that interventions that randomly rewire edges would increase spread. However, for other contagion models that require multiple exposures before adoption (i.e. complex contagions), recent work has argued for the opposite conclusion: highly clustered, rather than random, networks facilitate spread. Here we show that slight modifications of prior analyses, which make them more realistic, reverse this result. 7 - Two Stage Stochastic P-Order Conic Mixed Integer Programs Tight Second Stage Formulations Yingqiu Zhang, Virginia Tech, Blacksburg, VA, 24060, United States We study two-stage stochastic p-order conic mixed integer programs (TSS-CMIPs) in which the second-stage problems have p-order conic constraints along with integer variables. We provide sufficient conditions under which the integrality constraints on the second-stage integer variables of the TSS-CMIP can be relaxed, without effecting the integrality of the optimal solution of the problem, by adding parametric (non)-linear inequalities a priori. We perform computational experiments to evaluate the effectiveness of this second-stage convexication approach for solving TSS-CMIPs with second-order conic constraints in the second stage. 8 - Operations Research for Blood Donor Management Nico M. Van Dijk, University of Twente, Enschede, Ph.D.7522 NB, Netherlands Blood Management is of general societal interest. Dutch blood donations involve 750,000 yearly voluntary donations. By three projects: on blood inventory, on donor delays at collection sites, and on selective donor recruitment, it will be highlighted how Operations Research can provide quantitative and formal support. The results are real-life based on Dutch Blood Supply. 9 - Political Districting with Fairness Objectives: An Optimization-based Framework Rahul Swamy, University of Illinois at Urbana-Champaign, Urbana, IL, 61801, United States, Douglas M. King, Sheldon H. Jacobson Political districting (PD) is a problem of national interest. Classical models for PD focus on non-political objectives such as compactness. This paper addresses the question: How can voter information (e.g. polling data) be used to find politically fair districts Three such criteria are used, based on fundamental fairness principles such as proportionality (efficiency gap), partisan (a)symmetry, and competitiveness. A multilevel algorithm reduces instance sizes by graph contraction and solves an exact nonlinear bi-objective problem. A case study in Wisconsin shows Pareto-frontiers between the objectives, and that solutions that are politically fair are still reasonably compact.

344