Informs Annual Meeting 2017

MB49

INFORMS Houston – 2017

MB49

2 - Equilibrium and Ridesharing Payment Policies for Heterogeneous Dynamic Ridesharing Travelers in Morning Commute Rui Ma, University of California, Davis, 2041 Academic Surge, University of California, One Shields Avenue, Davis, CA, 95616, United States, drma@ucdavis.edu, Michael Zhang We formulate a dynamic ridesharing problem for a single bottleneck in morning commute. Travelers’ choices of departure-time and ridesharing modes as groups with heterogeneous values of travel time are captured at the dynamic equilibrium. An equivalent differential complementarity system is then established and solved numerically. Dynamic parking pricing strategies to achieve the system optimum at the free-flowing condition are discussed. The results suggest that the ridesharing payment policies should be properly designed to achieve the social, economic and environmental goals. 3 - Recognizing Taxi Driver Search Strategies using Sparse Subspace Clustering Kenan Zhang, Northwestern University, Evanston, IL, United States, KenanZhang2020@u.northwestern.edu, Yu Nie The rapid expansion of Transportation Network Companies (TNC) has revived the interest in understanding the search behaviors of traditional taxi drivers. Using a taxi GPS trajectory dataset collected in Shenzhen, China, this study explores taxi drivers’ search strategies in the urban area. Customer search behaviors are represented by features including searching distance and the characteristics of origin and destinations. A sparse subspace clustering (SSC) model is applied to classify taxi driver search strategies into major groups. Compared to conventional clustering algorithms (e.g. K-means), SSC better addresses the issue of nonlinear and high-dimensional nature of human behavior. 4 - Equilibrium Analysis of Urban Road Networks with Ride-sourcing Services The surge of ride-sourcing services in recent years has reshaped travelers’ mode choices for intra-city trips. As outcomes of the interactions between service operators and travelers, a substantial portion of vehicular trips in ride-sourcing services are spent searching and meeting travelers other than fulfilling their requested orders. These accessory trips, however, are missing ingredients in the current network equilibrium analysis methodology. Therefore, this study aims to propose a new model that describes the mixed equilibrium between the occupied vehicular trips that directly serve passenger demands and the accessory trips made by vacant ride-sourcing vehicles. 5 - Identifying Malevolent Users in Free-floating Bike Sharing Systems Aritra Pal, USF, 2210 Fitness Club Way, Apt 202, Tampa, FL, 33612, United States, aritra1@mail.usf.edu, Yu Zhang The purpose of this study is to extract and identify users who are deliberately mishandling or damaging bikes in a (free-floating) bike sharing system based on its historical data. We propose a four step method to solve this problem and illustrate its usage with “Share-A-Bull Bikes” free-floating bike sharing system as our case study. Zhengtian Xu, University of Michigan, Ann Arbor, MI, 48105-2540, United States, xzt@umich.edu, Yafeng Yin 361D Economics of Information Goods Sponsored: Information Systems Sponsored Session Chair: Zhe Zhang, University of Texas at Dallas, jameszhangzhe@utdallas.edu 1 - Towards Better Social Advertising Shaojie Tang, UT Dallas, Richardson, TX, United States, sxt146930@utdallas.edu Social advertising has the potential to general a significant cascade of adoption through influence in the online social networks. The goal of this work is to optimize the ad allocation such that the expected cascade can be maximized. 2 - Equilibrium Distribution of Mobile App Quality and Privacy Jiali Zhou, Hong Kong University of Science and Technology, Hong Kong, Hong Kong, zhoujiali90@163.com, Kai Lung Hui Mobile APPs bring consumers great convenience but raise privacy concerns because of the collection and use of customer data. We examine the equilibrium distribution of mobile APP quality and data requests. Consumers have heterogeneous preferences for APP quality and privacy. APP developers strategically choose quality and data requests to maximize profits. Our theoretical analysis suggests that the equilibrium data requests and hence consumer privacy follow a power law distribution. APP quality decreases monotonically, with high quality APPs requesting more personal data but also earning more market share. Preliminary empirical evidence from leading APP platforms support our analysis. MB51

361B Behavioural OR in Healthcare Invited: InvitedBehavioral Aspects of OR Invited Session Chair: Sally C. Brailsford, University of Southampton, Southampton, S017 1BJ, United Kingdom, s.c.brailsford@soton.ac.uk Co-Chair: Paul Harper, Cardiff University, United Kingdom, harper@cardiff.ac.uk 1 - Overdiagnosis Risk in Breast Cancer Screening: Incorporating Adherence Behavior and Racial Disparity Mahboubeh Madadi, Assistant Professor, Louisiana Tech University, Ruston, LA, United States, madadi@latech.edu, Wesley Wilkes Although known to be the primary method for early detection of breast cancer, mammography screening exposes patients to the risk of overdiagnosis, which is defined as the detection and subsequent treatment of a breast cancer that would not present clinically in a patient’s lifetime. In this study, we develop a Markovian model to estimate the risk of overdiagnosis under various screening policies while incorporating patient’s adherence behavior. We also provide personalized estimations for overdiagnosis risk taking into consideration different factors such as racial disparity in cancer incidence and progression as well as patient’s age. 2 - Operating Theatres Scheduling Optimization by Integrating Efficiency with Patients Satisfaction Xue Yang, Xi’an Jiaotong University, Xi’an, China, yxueyx@outlook.com, Wentian Cui, Zhen Shang, Yang Wang, Yang Zhang We building a mixed integer programming model for the operating theatres scheduling problem in a specialized hospital in China. In this model, we use a multinomial logit choice model to characterize patients satisfaction. The efficiency of operating theatres is measured by the linear combination of the number of finished surgeries and overtime. From extensive experimental analysis, we find interesting phenomena and propose valuable suggestions to healthcare managers. 3 - Are Patients Rational Decision-makers? In traditional OR models, it is typically assumed that there is a single, rational decision-maker, or a group of rational decision-makers who seek a compromise solution in the absence of a solution that is optimal for all. For example, health economists seek to maximize quality adjusted life years, and public health physicians (and insurance companies) take a population-level perspective. But how does the individual patient, or their family, take decisions regarding their own health? Are patients always rational, and if not, how can we model their behavior? 361C Modeling and Analysis of Emerging Mobility Services III Sponsored: TSL, Urban Transportation Sponsored Session Chair: Yang Liu, National University of Singapore, Singapore, 117576, Singapore, iseliuy@nus.edu.sg Co-Chair: Yu Nie, Northwestern University, Evanston, IL, 60208, United States, y-nie@northwestern.edu 1 - Multi-cycle Optimal Taxi Routing with E-hailing Xinlian Yu, University of Massachusetts Amherst, 130 Natural Sally C.Brailsford, University of Southampton, School of Management, Southampton, S017 1BJ, United Kingdom, s.c.brailsford@soton.ac.uk MB50

Resources Road, Amherst, MA, 01003, United States, xinlianyu@umass.edu, Song Gao, Hyoshin (John) Park, Xianbiao Hu

An optimal taxi routing problem is investigated for a single taxi that accounts for multiple cycles of pick-up and drop-off into the future. Practical implementations are proposed to solve the multi-cycle optimal taxi routing problem in a reasonable time. The solution will be compared with observed searching behaviors from GPS trajectories in a mega city to demonstrate the advantage of the multi-cycle approach.

182