Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

SB77

polyhedral analysis to reformulate a bilevel model as a single stage equivalent with facet defining inequalities. Results show that locating a station can counterintuitively decrease system effectiveness. We find that, even in the worst scenario, optimally locating stations can dramatically increase system performance and provide significant gains. 2 - Benders Decomposition for Profit Maximizing Hub Location Problems with Capacity Allocation Sibel Alumur Alev, University of Waterloo, 200 University Avenue West, Waterloo, ON, N2L 3G1, Canada, Gita Taherkhani, Seyed Mojtaba Hosseini We model capacity allocation decisions within profit maximizing hub location problems to satisfy demand of commodities from different market segments. We present a deterministic formulation of this problem and further extend this model considering stochastic demands. We describe two exact algorithms based on a Benders reformulation to solve large-size instances of the problem. Stochastic problem is solved using sample average approximation. Computational results show that large-scale instances can be solved to optimality, and that the proposed Benders decomposition algorithms generate cuts that outperform the best known cuts. 3 - GIS and the Subtle Proliferation of Optimization This paper discusses optimization and its formalization to support geographic analysis, planning, management and understanding. It is highlighted that GIS often represent a starting (or ending) point for undertaking optimization through location analytics capabilities. This paper provides an overview of representative models that can be found, highlighting both strengths and weaknesses, including issues of critical evaluation in observed performance and behavior of associated spatial optimization approaches in practice. 4 - Weber: “Should We Forget About Him in Location Science?” Richard Church, University of California, Santa Barbara, 1832 Ellison Hall, UCSB, Santa Barbara, CA, 93106-4060, United States Many now recognize that Launhardt’s location triangle has eclipsed that of Weber’s. Perreur (1998) actually asked the question in his paper: “should Weber be forgotten”? In this paper, I will demonstrate that Weber actually proposed a family of location problems that has yet to be tackled. 5 - Integrated Location and Inventory Modeling for Service Systems: Condition-based Replacements with IoT-based Data Erhan Kutanoglu, University of Texas-Austin, OR/IE Graduate Program, Dept of Mechanical Engineering, Austin, TX, 78712, United States, Murat Karatas We analyze a joint location and inventory problem in service parts logistics taking advantage of data on equipment condition via Internet of Things (IoT) technology. Instead of using a traditional replace-upon-failure policy, our model uses a condition-based part replacement (CBR) policy, finding the optimum condition to replace the parts, while designing the network and allocating the base stock levels. Even though CBR generates more frequent demands which put pressure on inventory, we capture the ideal balance between facility, transportation, inventory, and downtime costs, overall leading to significant cost savings in our preliminary results. n SC01 North Bldg 121A Joint Session OPT/Practice Curated: Applications of Stochastic Mixed Integer Programming Sponsored: Optimization/Optimization Under Uncertainty Sponsored Session Chair: Haoxiang Yang, Northwestern University, Evanston, IL, 60208, United States 1 - Optimizing Crashing Decisions in a Project Management Problem with Disruptions Haoxiang Yang, Northwestern University, 2145 Sheridan Road, Room C151, Evanston, IL, 60208, United States, David Morton We introduce a general type of sequential decision problem under uncertainty, where the uncertainty consists of a small number of disruptions, both the magnitude and the timing of which can be random. We consider a special case: a project crashing problem under a single disruption. When a disruption occurs, the duration of an activity which has not started could change. The magnitude of the activity duration change and the timing of the disruption can be random. We formulate a stochastic mixed integer programming (SMIP) model with second stage a mixed integer program. We propose a learning-based branch-and-bound algorithm and evaluate the computational performance of our approach. Sunday, 1:30PM - 3:00PM Alan Murray, University of California at Santa Barbara, Department of Geography, Santa Barbara, CA, 93106, United States

n SB77 West Bldg 213A Data Driven Insights for Decision Making Sponsored: Public Sector OR Sponsored Session Chair: Tina Rezvanian, Northeastern University, Boston, MA, 02115- 4420, United States Co-Chair: Ergun Ozlem, Northeastern University, Boston, MA 1 - The Foodome: Systematic Longitudinal Evaluation of Dietary Factors Associated with Coronary Heart Disease Soodabeh Milanlouei, Northeastern University, Boston, MA, 02130, United States Coronary Heart Disease is mainly driven by environmental factors among which diet is a key player. Nutritional epidemiology aims to explain the effect of nutrients on the heart by focusing on one nutrient at a time. Limited to a panel of 150 nutrients, epidemiologists are unable to address the role of a wide range of smaller molecules that we consume through diet. We propose a new methodology called Chemical-Wide Association Study (CWAS), which systematically evaluates multiple chemical factors simultaneously in relation to diseases in an agnostic fashion. We show that CWAS is not only capable of reproducing prior knowledge on diet-disease associations but also detecting new associations. 2 - Demand Forecast of Emergency Response Operations at United Nations High Commissioner for Refugees Tina Rezvanian, Northeastern University, Boston, MA, 02115- 4420, United States United Nations High Commissioner for Refugees runs Emergency Response supply chain for unprecedented demand by 65.6 million forcibly displaced people worldwide. Efficient utilization of available funds significantly depends on accurate regional demand estimations. Although complex models incorporate higher degrees of data characteristics, we find that models in humanitarian logistics need to capture the right balance between flexibility and practicality. This paper aims for optimizing the use of techniques and providing a framework that promotes practicality of advanced analytics 3 - Donations for the Refugee Crisis: Cash Versus In-kind Assistance Telesilla Oly Kotsi, PhD Candidate, Indiana University, Bloomington, IN, United States, Owen Wu, Alfonso J. Pedraza-Martinez Our study is motivated by fieldwork during the response to the Syrian refugee crisis in north-western Greece. We analyze tradeoffs regarding the mix of assistance (cash and in-kind assistance) that a humanitarian organization should offer to refugee populations. We model a humanitarian organization that aims to maximize its benefit to refugees, local retailers, and local residents. We determine the impact that cash assistance has on all above actors. We find that the introduction of cash assistance increases the importance of the allocation decision of the remaining budget of the humanitarian organization among the goods provided as in-kind assistance to refugees. 4 - Improving Financial Loan Access for Farmers in Uganda Erica Gralla, Assistant Professor, George Washington University, Washington, DC, United States, Jillian Miles Ugandan farmers struggle to access financial loans. We aimed to understand the dynamics driving this problem and identify the most effective interventions to increase access to loans. A system dynamics model was developed to integrate data from diverse sources. The results highlight some of the most significant barriers to loan access — interest rates and geographic access to financial institutions — and identify combinations of interventions that can lead to broader loan access. Location and Logistics Sponsored: Location Analysis Sponsored Session Chair: Sibel Alumur Alev, University of Waterloo, Waterloo, ON, N2L 3G1, Canada 1 - Locating Law-enforcement Facilities to Intercept Evading Vehicle Flows Okan Arslan, HEC Montreal, 5308 Ave Decelles, Montreal, QC, H3T1V8, Canada, Ola Jabali, Gilbert Laporte We consider locating law enforcement facilities in a transportation network to intercept unlawful vehicle flows traveling between origin-destination pairs, who in turn deviate from their routes to avoid any encounter with such facilities. Vehicle deviations are bounded by a given tolerance. We use duality theory and n SB78 West Bldg 213B

60