Informs Annual Meeting 2017

SC41

INFORMS Houston – 2017

SC40

3 - Interactive Evolutionary Multiobjective Optimization Driven by Decision Rules Roman Slowinski, Professor, Poznan University of Technology, Pl. Marii Sklodowskiej-Curie 5, NIP: 777-00-03-699, Pozna , 60-965, Poland, roman.slowinski@cs.put.poznan.pl, Salvatore Corrente, Salvatore Greco, Salvatore Greco, Benedetto Matarazzo Evolutionary multiobjective optimization can lead more quickly to a small subset of most preferred non-dominated solutions if the search of the solution space is driven by a preference model constructed simultaneously with the search. Here we propose a preference model in terms of “if , then ” decision rules induced from rough approximations of preference information given by the user every k generations in terms of classification of a small sample of solutions into “promising” and “to-be-avoided”. We present results of very encouraging experiments with this method performed on some well-known benchmark problems. 4 - A Multicriteria Approach to Support the Performance Evaluation Process for Tariff Revisions in Brazil Adiel Teixeira De Almeida Filho, Universidade Federal de Pernambuco, Caixa Postal 7471, Cx Postal 7471, Av. General San Martin,1083, Recife-PE, 50.630-971, Brazil, atalmeidafilho@yahoo.com.br, Danielle Morais, Luanna T. da Silva Lira, José R.Figueira, Ana Paula C. Costa Electrical power tariffs in Brazil may be revised according to recent changes introduced by a law approved in 2013. The agency that regulates the Brazilian electric system considers a decision model that includes a benchmark evaluation based on a DEA model to estimate operating costs, followed by public hearings to decide upon a tariff revision for each electricity company. Such decision process is relevant once it affects the profitability of electricity companies as well as final customers’ electricity bill. This work presents the ongoing research about this subject and how a multicriteria decision-making technique may contribute to this decision process, which is currently based only on DEA. Sponsored: Analytics Sponsored Session Chair: Harrison Schramm, CANA Advisors, Pacific Grove, CA, 93950, United States, harrison.schramm@gmail.com 1 - Microservice Architecture for Delivering Analytic Models Jon A.Higbie, Revenue Analytics, 3100 Cumberland Blvd., Suite 1000, Atlanta, GA, 30339, United States, jhigbie@revenueanalytics.com Since 2011 Micro Service Architecture (MSA) has been gaining favor as an architecture and method for developing software, particularly in the cloud. MSA breaks an application down to small components that perform a single function. These microservices then communicate with each other via web services. If done correctly, the resulting application has high reliability and the microservices are easily reusable. Learn how this approach is changing the way complex analytic solutions are delivered. 2 - Hetergenous Quantum Computing Gideon Bass, Booz Allen Hamilton, Washington, DC, United States, tb Quantum computing offers the possibility of producing significant speed-up and improved solution quality when compared to traditional computing. There are now commercially available quantum annealing (QA) devices that are designed to solve difficult optimization problems. We present a heterogeneous computing stack that combines QA with traditional hardware and allows usage of QA on problems larger than existing quantum hardware could solve in isolation, and, as a representative problem, we discuss a sample satellite constellation assignment problem, and the exciting results of this research. SC39 352D Emerging Concepts

352E Doing Good with Good OR Invited: Doing Good with Good OR Invited Session Chair: Turgay Ayer, Georgia Institute of Technology, School of Industrial and Systems Engineering, Groseclose 417, Atlanta, GA, 30332, United States, ayer@isye.gatech.edu Co-Chair: Jonathan Helm, Indiana University, Kelley School of Business, 1309 East Tenth Streeet, Bloomington, IN, 47405, United States, helmj@indiana.edu 1 - Doing Good with Good OR - Student Paper Competition Turgay Ayer, Georgia Institute of Technology, School of Industrial and Systems Engineering, Groseclose 417, Atlanta, GA, 30332, United States, ayer@isye.gatech.edu Doing Good with Good OR - Student Paper Competition is held each year to identify and honor outstanding projects in the field of operations research and the management sciences conducted by a student or student group that have a significant societal impact. 2 - Optimal Design ofEfficient Rooftop Photovoltaic Arrays Madeleine Udell, Cornell University, 301 Ithaca Road, New York, Ithaca, NY, 14850, United States, madeleine.udell@gmail.com Abstract not available. 3 - Mobile Money Agent Inventory Management Karthik Balasubramanian, Harvard Business School, 25 Harvard Way, Morgan T81, Boston, MA, 02163, United States, kartbala@gmail.com Mobile money agents exchange cash for electronic value and vice versa, forming the backbone of an emerging electronic currency ecosystem in the developing world. We model the agent’s inventory problem, develop policy recommendations, and evaluate these policies with East African data. 4 - Truthful Mechanisms for Medical Surplus Product Allocation Can Zhang, Georgia Institute of Technology, 499 Northside Circle

NW, Apt 315, Atlanta, GA, 30309, United States, czhang2012@gatech.edu, Atalay Atasu, Turgay Ayer, Beril L. Toktay

We analyze resource allocation problems faced by medical surplus recovery organizations (MSROs) that recover reusable medical products to fulfill the healthcare needs of under-served regions/countries. We present strategies to improve MSROs’ value provision when recipient needs are private information, and validate our results using historical data from a partner MSRO.

SC41

352F Health Care, Modeling and Optimization Contributed Session Chair: Shanshan Wang, Beijing Institute of Technology, Beijing, China, shshwang_bit@163.com 1 - Risk-averse Inpatient Discharge Planning Maryam Khatami, Texas A&M.University, 4050 ETB, College Station, TX, 77840, United States, maryam.khatami@tamu.edu, Lewis Ntaimo, Mark Alan Lawley We study the inpatient discharge planning (IDP) problem on a daily basis. Discharge delay deteriorates patient satisfaction, incurs extra costs to the hospital and additional stress to the staff, and causes admission delay for patients in upstream units. We formulate the IDP problem as a risk-averse two-stage stochastic program. Our model captures two sources of uncertainty: inpatient discharge processing time and bed request time. To the best of our knowledge, we are the first to apply the stochastic decomposition algorithm to solve a risk-averse stochastic model. We solve our model for several instances of the IDP problem generated using data from a Texas hospital.

81