Informs Annual Meeting 2017

SB49

INFORMS Houston – 2017

SB49

3 - Optimizing Paratransit Operations to Serve Demand of Passengers with Disabilities in Urban Areas Eric Gonzales, University of Massachusetts-Amherst, 130 Natural Resources Way, Civil and Environmental Eng, Amherst, MA, 01003, United States, gonzales@umass.edu, Mahour Rahimi Demand responsive transit service for people with disabilities is typically provided in the United States with paratransit services managed by transit agencies, as required by the Americans with Disabilities Act (ADA). The cost of operating a dedicated fleet of vans and cars for these customers is rising as more people become eligible for ADA service. Demand and operations data for NJ Transit support models and analysis to optimize zoning strategies, garage locations, and transfer policies. An opportunity now exists to further improve costs and mobility by coordinating ADA service with transportation network companies, which has been piloted by the MBTA in Boston. 4 - Inferring Origin-Destination Demand Matrix and Utility-Based Travel Preferences in Multi-Modal Travel Environment using Automatic Fare Collection Data Laiyun Wu, SUNY-Buffalo, 326 Bell Hall, Buffalo, NY, 14226, United States, laiyunwu@buffalo.edu We develop inference methods for understanding and expressing public transit system utilization based on travelers’ origins/destination (OD) and ensuing behavior. The presented methodology infers both the preference vector and true OD-pair demand by analyzing Automatic Fare Collection (AFC) data, as travelers make their multi-modal route choice decisions in the time-dependent travel environment. 361D Digital Advertising Sponsored: Information Systems Sponsored Session Chair: Manmohan Aseri, The University of Texas at Dallas, Richardson, TX, 75080, United States, manmohan.aseri@utdallas.edu 1 - Block the Ad-block Users? Not Always Manmohan Aseri, University of Texas-Dallas, 2200 Waterview Parkway, Apt 1738, Richardson, TX, 75080, United States, manmohan.aseri@utdallas.edu, Milind Dawande, Ganesh Janakiraman, Vijay S.Mookerjee Ad-Blockers are the software or the plug-ins which remove ads from the users’ browsers when they access a website on the internet. Companies which are heavily dependent on the advertisement revenue dislike the ad-blocker usage due to the loss of the ad revenue. We address the problem from the perspective of those publishers (websites) which rely heavily on the ad revenue and provide insights about how these companies should react to ad-block users. We solve the problem using a game theoretic model. Our results suggest that the network effect plays a critical role in the reaction of a publisher to the ad-block users. These results are consistent with the current practice. 2 - When are Two Machines Better than One? A Model to Reduce Rental Cost in Cloud Computing Leila Hosseini, Information Systems, The University of Texas at Amazon Web Services (AWS), a major cloud provider in the cloud infrastructure market, offers a diverse set of computing resources called virtual machines (VM). All VMs are priced based on their capabilities such that VMs whose prices are higher perform better than those which have lower prices. In this setting, AWS users receive a lot of flexibility to select their desired VMs. However, it is difficult to encounter to this much flexibility without being confused at the same time. Therefore, it is a must for users to find an efficient solution to transform this confusing market into a beneficial one. We study AWS cloud user’s problem of renting VMs to minimize rental cost in the presence of a time constraint. 3 - Measuring Channel Complementarities in an Online Advertising Supply Chain Changseung Yoo, The University of Texas at Austin, Austin, TX, 78712, United States, csyoo@utexas.edu, Anitesh Barua, Genaro J. Gutierrez In the empirical literature on testing complementarities, no distinction has been made between the existence of complementarity and recognizing and acting upon such complementarity. To address this gap, we analyze how a decision maker’s understanding of complementarities can affect the empirical evidence of complementarities or substitutabilities within a structural model. SB51 Dallas, Richardson, TX, 75080, United States, leila.hosseini@utdallas.edu, Vijay S.Mookerjee, Chelliah Sriskandarajah

361B Behavioral Collaboration Invited: InvitedBehavioral Aspects of OR Invited Session Chair: Leroy White, Professor, Warwick Business School, Warwick Business School, United Kingdom, Leroy.White@wbs.ac.uk 1 - The Problems of Collaborative Behaviour and Operations Research: An Overview Leroy White, University of Bristol, School of Eonomics Finance and Accounting, 12 Priory Road, Bristol, BS8 1TU, United Kingdom, leroy.white@bristol.ac.uk Abstract not available. 2 - Affordances for Collaborative Behavior Operations Research Interventions: A Practice-based View Katharina Burger, University of Portsmouth, Business School, OSM, Portland Street, Portsmouth, PO1 3DE, United Kingdom, katharina.burger@port.ac.uk Theorizing from practice-based OR interventions is a critical task for behavioral operations researchers who aim to understand how collaborative performance can be scaffolded. A micro-level lens to realize theory-based empirical studies of affordances is proposed. Drawing on data from a low-tech facilitated modelling workshop, three key dimensions of behavior with OR performance scaffolds are highlighted: trust, entrenchment and sharedness. 3 - Building the Multiplex: An Agent-Based Model of Formal and Informal Network Relations Duncan Robertson, Loughborough University, SBE, Loughborough, United Kingdom, d.a.robertson@lboro.ac.uk, Leroy White We present a new agent-based model of group decision making, where links are made in hierarchical organizational structures. Our model presents a trade off between deliberate organizational hierarchical planning and allowing organizational ‘short cuts’ to form spontaneously. We present results showing when organizations should be planned and when emergent designs can offer higher performance. 361C Large-scale Data Analytics in Urban Transportation Modeling II Sponsored: TSL, Urban Transportation Sponsored Session Chair: Shanjiang Zhu, George Mason University, Fairfax, VA, 22030, United States, szhu3@gmu.edu Co-Chair: Zhen Qian, Carnegie Mellon University, Carnegie Mellon University, Pittsburgh, PA, 15213, United States, seanqian@cmu.edu 1 - Statistical Inference of Probabilistic Origin-destination Demand using Day-to-day Traffic Data Zhen Qian, Carnegie Mellon University, 4800 Forbes Avenue, Pittsburgh, PA, 15213, United States, seanqian@cmu.edu, Wei Ma Our research develops a novel framework for estimating the mean and variance/covariance of Origin-Destination demands considering the route choice variation and statistical equilibrium simultaneously. The observability and non- uniqueness property of the estimated probabilistic O-D demand are discussed. The framework is evaluated in a small and a large-scale network to provide insights and to demonstrate computational efficiency. 2 - How Does Social Media Infer the Longitudinal Travel Behavior? Qing He, University at Buffalo, The State University of New York, 313 Bell Hall, Buffalo, NY, 14260, United States, qinghe@buffalo.edu, Zhenhua Zhang, Shanjiang Zhu This paper proposes a sequential model-based clustering method to group the high-resolution Twitter locations and extract the Twitter displacements. Further, this study compares the displacement-based results from Twitter data with the results from traditional household travel survey. Our study demonstrates the potentials of employing social media to infer longitudinal travel behavior and to support travel demand modeling for a metropolitan area. SB50

52