Informs Annual Meeting 2017

TD09

INFORMS Houston – 2017

TD09

2 - Continuous Location Model for Mobile Fire Fighting Resources Pablo Escalona, Universidad Técnica Federico Santa María, Av Espana 1680, Valparaíso, 2390123, Chile, pablo.escalona@usm.cl, Monica Lopez-Campos, Mario Navarrete The Valparaiso region (Chile) has suffered more than 16,000 fires between 1998 and 2015, including a year surpassing 1,300 fires (CONAF, 2015). For its special topographic characteristics, air resources have a high degree of significance. A modified version of the “Queuing Location Problems on the Plane” model (Drezner et al., 1990) has been selected to determine the optimal location of airbases, in an attempt to minimize the amount of burned hectares by wildfires. The computational study determines: the optimum location of the base, which parameters have greater influence in the optimum location, and which parameters have greater influence in the computational times. 3 - A Resource Allocation Model for Large Scale Bio-security Events Lauren Gardner, Senior Lecturer, University of New South Wales, Kensington Campus, Building H20, Sydney, 2052, Australia, l.gardner@unsw.edu.au, David Rey Real-time strategies for control are required to minimize the spread of newly emerging infectious diseases at a national and international level. Currently, there is a gap in the literature for optimization-based decision control models that account for the heterogeneous nature of the air travel patterns and infectious disease spreading dynamics. In this work a bi-level optimization model is proposed where the objective is to minimize the risk of a contagion spreading via the air traffic network, where the decision is the allocation of control resources to airports during an on-going event. 4 - Modelling Ambulance Dispatch Systems During Extreme Weather Events Eric B. DuBois, University of Wisconsin-Madison, Industrial & Systems Engineering, 1513 University Avenue, Madison, WI, 53706, United States, edubois2@wisc.edu, Laura Albert A moderate, but prolonged increase in patient arrivals, as often occurs during severe weather events, can overwhelm emergency responders and lead to patient queuing and patient health deterioration. To examine this issue we develop a Markov decision process to determine a policy for withholding ambulances from less serious patients in the expectation of more emergent future calls. We further develop heuristics, which capture most of the benefit of the optimal policy while allowing for easier implementation by dispatchers. The results are illustrated with computational examples. 5 - Optimizing Healthcare Facility Locations to Intervene a Fast Spreading Epidemic Gonca Yildirim, Asst. Prof., Cankaya University, Eskisehir Yolu 29 km Yukariyurtcu Mah., Mimar Sinan Cad. No 4, Ankara, 06790, Turkey, goncayildirim@cankaya.edu.tr, Ayyuce Aydemir-Karadag Emerging and re-emerging epidemic diseases such as Ebola continue to threaten the global health security. During an outbreak, timely case management is crucial to interrupt chains of transmission and prevent further spread of the infection. The most effective intervention policies are to remove infected individuals from the population by quarantine and provide treatment in healthcare facilities. We consider a fast-spreading epidemic with a short incubation period and take into account forecast models developed for the progression of the outbreak in an environment with scarce resources in an integer programming model to determine locations of healthcare facilities dynamically over time. 6 - Location and Allocation of EMS Vehicles in Dougherty County, Albany, GA Damitha Bandara, Assistant Professor, Albany State University, 504 College Drive, Albany, GA, 31705, United States, damitha.bandara@asurams.edu Dougherty County EMS is an advanced life support ambulance service that provides quality pre-hospital medical care and transportation to the citizens of Dougherty County. The contribution of this research towards improving the efficiency of the Dougherty County EMS system is twofold. One is to determine the optimal locations for EMS system stations and the other is to determine the response territory for each EMS vehicle. A mathematical model will be developed to obtain the optimal locations and response boundaries of EMS system in Dougherty County. Computational results will be used to compare the performance of the proposed EMS system as opposed to the existing system.

330A Learning, Information, and Games Sponsored: Manufacturing & Service Oper Mgmt Sponsored Session Chair: Kostas Bimpikis, Stanford University, 655 Knight Way, Stanford, CA, 94305, United States, kostasb@stanford.edu 1 - Learning in Repeated Auctions with Budgets: Regret Minimization and Equilibrium Yonatan Gur, Stanford University, Knight Management Center, 655 Knight Way, Stanford, CA, 94305, United States, ygur@stanford.edu, Santiago Balseiro We study how budget-constrained advertisers may compete in sequential auctions for available ad slots in the presence of uncertainty about future opportunities and competitors’ heterogeneous preferences, budgets, and strategies. We introduce a family of dynamic bidding strategies we refer to as adaptive pacing strategies, in which advertisers adjust their bids throughout the campaign according to the sample path of expenditures. Under arbitrary competitors’ bids, we establish through matching lower and upper bounds the asymptotic optimality of this class of strategies. Moreover, we characterize a regime under which these Xiaosheng Mu, Harvard University, Cambridge, MA, United States, xiaoshengmu@fas.harvard.edu, Annie Liang, Vasilis Syrgkanis Consider a decision-maker who sequentially acquires Gaussian signals from heterogeneous sources for a future decision. Which sources should he choose to observe in each period? In environments that we characterize, it turns out that the dynamically optimal path of signal acquisitions: (1) exactly coincides at every period with the myopic path of signal acquisitions, and (2) achieves ``total optimality,” so that at every late period, the decision-maker will not want to revise his previous signal acquisitions even if given this opportunity. We show that generically, these properties hold at all sufficiently late times, so that the dynamically optimal path and myopic path are eventually equivalent. 3 - Learning and Efficiency in Games with Dynamic Population Thodoris Lykouris, Cornell University, 107 Hoy Rd, Ithaca, NY, United States, teddlyk@cs.cornell.edu, Vasilis Syrgkanis, Eva Tardos We study multi-player game settings (online advertising, internet routing, and bandwidth allocation) where the player set is dynamically evolving over time and participants apply online learning algorithms to adapt to the changing environment. Traditional equilibrium notions require too much information for the players (since they need to form perfect beliefs about the identity and the behavior of other participants) to extend to dynamic settings. In contrast, we show that, in a broad class of games, learning agents that can adapt to the changing environment reach outcomes of high social welfare even when there is a high turnover in the population, extending traditional price of anarchy bounds. 4 - Information Aggregation in Heterogenous Markets Yaarit Even, Columbia University, 25 Claremont Ave #1c, New York, NY, 10027, United States, yeven18@gsb.columbia.edu, Alireza Tahbaz-Salehi, Xavier Vives This paper argues that the presence of heterogeneity in traders’ trading costs/risk aversion can reduce price informativeness and result in an informational externality. We show that in the presence of heterogeneity, the information content of the price is biased towards the private signals of agents with lower trading costs, thus reducing the informativeness of the price. More importantly, we show that such heterogeneity results in an informational externality with first-order welfare implications. strategies approximate Nash equilibrium in dynamic strategies. 2 - Optimal Learning from Multiple Information Sources

378