Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

WB78

n WB76 West Bldg 212C Navigating the Funding Landscape Sponsored: Minority Issues Sponsored Session Chair: Trilce Encarnacion, Rensselaer Polytechnic Institute, Troy, NY, 12180, United States 1 - Navigating the Funding Landscape Trilce Encarnacion, Rensselaer Polytechnic Institute, Troy, NY, 12180, United States This panel will explore a variety of opportunities for funding your ORMS research. Lessons learned and best practices to increase your funding success will be shared. Panelist Jennifer A. Pazour, Rensselaer Polytechnic Institute, Troy, NY, 12180, United States n WB77 West Bldg 213A Nonprofit Operations Management Sponsored: Public Sector OR Sponsored Session Chair: Gemma Berenguer Falguera, Purdue University, Purdue, West Lafayette, IN, United States 1 - Performance Metrics for Emergency Response Network Design Alexander Rothkopf, Massachusetts Institute of Technology, Cambridge, MA, 65197, United States, Jason Acimovic, Jarrod D. Goentzel We use a stochastic linear program to model a US-based emergency response organization’s network of warehouses to serve sudden onset disasters. We develop multiple performance measures to characterize the risk portfolio, the supply portfolio and the carrier portfolio to support decision-makers in evaluating the status-quo of their operations and suggest areas of improvement. We show how our model and the measures work with data from our case. 2 - Supply Constrained Location-distribution in Nonprofit Settings Gemma Berenguer, Purdue University, West Lafayette, IN, 47906, United States, Chong Park There is a need to design location-distribution problems in not-for-profit settings where limited resources have to be allocated to different demand regions and, thus, a combination of efficiency and equity goals can rightfully be considered. In this paper, we design and solve models that represent this setting. In particular, we propose the use of a well-known fractional efficiency measure and a new inequity measure related to the Gini coefficient that is based on the relative utility obtained by each demand region given a certain supply allocation. Our resolution technique allows us to use fractional objective measures that have not been commonly utilized before due to tractability issues. 3 - Value of Combining Patient and Provider Incentives in Humanitarian Healthcare Service Programs Karthik Natarajan, University of Minnesota, Minneapolis, MN, 55455, United States, Mili Mehrotra We analyze the incentive design problem faced by a budget-constrained humanitarian organization managing a healthcare delivery program. Incentives offered to the healthcare provider are aimed at improving the quality and quantity of services offered, and demand-side incentives are used to encourage patients to seek care. In practice, incentives are often targeted at improving the availability of healthcare services. However, our results suggest that by offering the right combination of incentives to the provider and patients, program performance can be improved significantly. Furthermore, using the right incentives is critical to fully realizing the benefits of fundraising efforts. 4 - Donor Funding for Drug Availability Iva P. Rashkova, Washington University in St Louis, St Louis, MO, United States Motivated by the Global Fund grant recipients, we study the procurement of health products subject to a donor budget constraint. Donor funding is in the form of lump-sum disbursements or per-unit subsidy agreements, or both. The donor objective is to maximize consumption in the face of demand uncertainty and grant recipients’ unobservable procurement costs. We derive the optimal allocation of donor funds, which exhibits a risk-hedging synergy between the two types of funding. Using Global Fund data, we find that implementing the optimal budget allocation includes both types of funding for 60% of grant recipients and could lead to up to 21% increase in total consumption.

n WB78 West Bldg 213B

Joint Session PSOR/TSL Urban: Sustainable Transportation/Logistics in Public Sector OR I Sponsored: Public Sector OR Sponsored Session Chair: Sung Hoon Chung, Binghamton University, Binghamton, NY, 13902, United States 1 - Setting up a Liquefied Natural Gas Refueling Infrastructure and Greenhouse Gas Emissions Savings on the Pennsylvania Turnpike Sang Jin Kweon, Carnegie Mellon University, Pittsburgh, PA, 15213, United States Recognizing the increased spotlight on liquefied natural gas (LNG) for long-haul fleets, logistics companies are interested in switching their long-haul fleets from diesel to LNG. In this talk, we present a bi-criteria binary linear programming model to locate LNG refueling stations on a directed transportation network with two conflicting objectives; maximizing the annual total vehicle-miles traveled covered by the stations and minimizing the capital cost for building the refueling infrastructure. We applied the proposed model to the Pennsylvania (PA) Turnpike System and estimate the potential reduction of pollutants on the turnpike as a function of the LNG refueling infrastructure cost. 2 - Automated Guidedway Transit (AGT) System Control Design with Q-learning Algorithm Young Jae Jang, KAIST, Daejeon, Korea, Republic of, Illheo Hwang The automated guideway transit (AGT) is a class of transportation system consisting of automated vehicles operating on a network of guideways or specially designed rails. This system is also know as Personal Rapid Transit (PRT) or pod- cars. One of the challenges in commercializing the AGT is the development of a routing algorithm controlling massive number of vehicles. We introduce AI based algorithm, Q-learning routing, which effectively control AGT system which consists of massive number of vehicles. Simulation analyses validate that the proposed algorithm outperforms the conventional DP based algorithms in terms of transit service times. 3 - A Hybrid Heuristic Method for the Zero Emission Vehicles (ZEV) Routing Problem with Fueling Stations Shichun Hu, University of Southern California, Los Angeles, CA, United States Recent researches on Vehicle Routing Problem (VRP) have included the effect of greenhouse gas emissions to help alleviate the global climate change. One direction is to incorporate ZEVs into VRP models. Despite their advantages of zero or near-zero emissions, the fact that ZEVs have less mile range per fueling and less fueling stations is likely to offset their benefits thus making it harder to replace diesel trucks. We propose a method that combines Adaptive Large Neighborhood Search with local search to solve the ZEV routing problem considering fueling stations. 4 - Optimal Location Design and Dynamic Pricing of Electric Vehicle Charging Facilities under Uncertain Demand and User Decisions Leila Hajibabai, Stony Brook University, Department of Civil Engineering, 2433 Computer Science, Stony Brook, NY, 11794, United States, Amir Mirheli This study develops a methodology that incorporates a dynamic pricing scheme into facility location design for electric vehicle charging infrastructure. The objective is to determine the optimal location and capacity of charging facilities in the transportation network and the optimal pricing based on demand. The problem is formulated as a bi-level optimization program that minimizes the total costs including infrastructure investments and operation costs of facilities under user-equilibrium flows. Numerical experiments confirm that the proposed methodology solves the problem efficiently.

469