2015 Informs Annual Meeting

POSTER SESSION

INFORMS Philadelphia – 2015

64 - Analytic Network Process: Assisting Computers to Think Like Humans? Elena Rokou, Chief Research Officer, Creative Decisions Foundation, Ellsworth Ave, Pittsburgh, PA, United States of America, erokou@gmail.com Whatever your stance is on Artificial intelligence, it is generally admitted that it has not yet enabled computers to make satisfactory decisions. Methods like Neural Networks, can train computers to make decisions for simpler types of tasks, but the ANP can factor in morality, ethics and broader considerations associated with complex decisions. We want computers to think more like humans, thoughtful and compassionate in their choices, and ANP enables this type of higher-level decision-making. 65 - An Energy-aware Multiobjective Scheduling Optimization Framework for Sustainable Manufacturing Saeed Rubaiee, Wichita State University, 2119 Malcolm Street, Wichita, KS, 67208, United States of America, ssal21@gmail.com, Mehmet Bayram Yildirim The goal of this paper is to minimize the total tardiness and total energy cost under time-of-use electricity tariffs, where energy prices vary hourly, on a non- preemptive single-machine. The problem is modeled using a mixed-integer multiobjective mathematical programming model to obtain an approximate Pareto front. Results show that the proposed multiobjective NSGA-II genetic algorithm finds a good approximate Pareto front with better diverse solutions and shorter computational CPU times. 66 - Early Warning Methods and Predictive Models for Hospital Risk and Readmissions Jakka Sairamesh, Ceo And President, CapsicoHealth, Inc, 2225 E Bayshore Rd STE 200, Palo Alto, CA, 94303, United States of America, ramesh@capsicohealth.com, Ruichen Rong This poster and research abstracts presents the effectiveness of methods for improving patient quality outcomes (e.g. reducing 30-day readmissions) based on clinical and cost based factors. We will present early-warning methods to predict patients at risk of 30-day readmissions based on past admissions, ER visit rates, mortality rates, and charges. The dominant factors includes clinical risk, costs, emergency room visits and mortality rates. The prediction showed nearly 88 percent accuracy. 67 - Software License Optimization Model for Software Asset Management Seungbae Sim, Korea Institute for Defense Analyses, 37 Hoegi-ro, Seoul, Korea, Republic of, sbsim@kida.re.kr, Cheonsoo Yoo Information System can be generally comprised of hardware and software. As software has been getting more important than hardware, most organizations must reduce increasing software costs and control software assets. Especially, commercial software can be licensed to end-users. We propose the mathematical model considering the complexity of software license types. Also, the case example is presented for validating the proposed optimization model. 68 - Optimization Problems Arising in Stability Analysis of Discrete Time Recurrent Neural Networks Jayant Singh, Dept. of Mathematics, North Dakota State University, 1210 Albrecht Boulevard Minard 408, Fargo, ND, 58102, United States of America, Jayant.Singh@ndsu.edu We consider the method of Reduction of Dissipativity Domain to prove global Lyapunov stability of Discrete Time Recurrent Neural Networks. It involves a multi-step procedure with maximization of special nonconvex functions over polytopes on every step. We derive conditions which guarantee an existence of at most one point of local maximum for such functions over every hyperplane. This nontrivial result is valid for wide range of neuron transfer functions. 69 - Modular Function Deployment Adapted to the Project Typology in the Development of Modular Products Monique Sonego, Universidade Federal do Rio Grande do Sul, Av. Osvaldo Aranha 99 - PPGEP 5 ∫ andar, 90035-190, Porto Alegre, Brazil, hgmonique@gmail.com, Angela Danilevicz, Márcia Echeveste Modular Function Deployment (MFD) is one of the best-known methods for modularization in New Product Development. However, this method is not tailored to different project typologies. We propose an adaptation for the MFD method for different levels of complexity and novelty of each project. This adaptation provides companies with the possibility of choosing the setting of stages and tools that best fit their specific projects by customizing the application of the MFD. 70 - Enriching Competitiveness and Connectivity with HLED-inspired Air Service Agreement Andrew Stapleton, Professor Of Supply Chain Management, University of Wisconsin La Crosse, 1725 State Street, La Crosse, WI 54650, La Crosse, WI, 54650, United States of America, Astapleton@uwlax.edu U.S. cargo and passenger airlines will have a greater opportunity to compete for a larger share of freight trade and traffic between the U.S. and Mexico when the new Air Services Agreement (ASA) takes effect January 2016. It is a key element

of the US-Mexico High Level Economic Dialogue (HLED), that aims to promote competitiveness and connectivity, foster economic growth, productivity and innovation, and partner for regional and global leadership. 71 - The Value of Flexibility in Dynamic Ride-sharing Mitja Stiglic, University of Ljubljana, Kardeljeva Ploöcad 17, We consider a dynamic ride-sharing system that allows people with similar itineraries and time schedules to share rides. Participants are willing to somewhat adapt their trip plans in order to be matched. We study how participants’ flexibility in departure times and the willingness of drivers to perform detours influence the matching rate and the sustainability of the system. We conduct an extensive computational study to quantify the impact on system performance in a variety of settings. 72 - Managing a Bike-sharing System using Wireless Mobility Data Rahul Swamy, University at Buffalo, 49 Englewood (Lower), Buffalo, NY, 14214, United States of America, rahulswa@buffalo.edu, Jose Walteros This research aims to provide a mathematical framework for operating a campus bike-sharing system. We use wireless network (WiFi) usage logs to generate a detailed estimation of the inter-building demand across campus. We propose solving a sequence of MILPs to determine: 1) the optimal location of bike stations, 2) the number of bikes to be added to or removed from each station every hour to satisfy the demand-supply needs, and 3) the redistribution logistics, while minimizing overall costs. 73 - Configurations of Distribution Strategies Based on 124 quantitative samples with both first-hand and second hand data, as well as 56 qualitative samples, this paper examines the strategic fit of distribution strategies from the perspective of configuration theory. We find that the fit between operational decisions including infrastructural and structural decisions, and operational competencies including cost and flexibility, has an important effect on business performance. 74 - Teaching Machine Learning Methods Based on Systematic Approach Derived from Potential Theory Nadia Udler, Fordham University, 113 West 60 St, New York, NY, United States of America, nadiakap@optonline.net Many real word problems can be reduced to black box optimization. One of the challenges in the design of black box optimization software is identifying a minimal set of modules for building hybrids for real word applications. Existing software provides such building blocks but they are heuristic thus difficult to teach. We discuss black box optimization library based on systematic approach derived from potential theory. It can be used as educational tool to teach machine learning techniques. 75 - Optimizing Player Lineups in Daily Fantasy Sports Nicholas Valentour, Graduate Student, University of Nebraska Omaha, Department of Mathematics, Omaha, NE, United States of America, nvalentour@gmail.com, Betty Love The growing market of online fantasy sports has increased demand for providers of daily player projections and optimal fantasy lineups. Fantasy lineup optimization is a variant of the multiple choice knapsack problem. We develop an integer linear programming algorithm to identify optimal daily lineups. Further, we combine the algorithm with forecasting to examine the overall fantasy performance on historical basketball data. 76 - Design and Operation of a Last Mile Transportation System Hai Wang, MIT ORC, 2D 550 Memorial Drive, Cambridge, MA, 02139, United States of America, haiwang@smu.edu.sg The Last Mile Problem refers to the provision of travel service from the nearest public transportation node to a home or office. Last Mile Transportation Systems (LMTS) are critical extensions to traditional public transit systems. We study the design and operation of a LMTS from three perspectives: (1) performance evaluation from a queueing perspective; (2) system operation from an optimization perspective; and (3) demand estimation from an inference perspective. 77 - Competition Strategies of Platform-based Retailing Man Wang, Guanghua School of Management, Peking University, No.5 Yiheyuan Road Haidian District, Beijing, China, dream26@pku.edu.cn, Lihua Chen While collaborating with third-party sellers via opening infrastructure online, platform-based retailers and third-party sellers run into a head-to-head price competition. We show that when the inventory of the platform-based retailer is sufficient, higher service quality can bring larger competitive profits. However, it may not always be optimal for a platform-based retailer to improve its service quality. The platform-based retailer may be worse off when the inventory is shortage. Jing Tang, Em-lyon Business School, 23 Avenue Guy de Collongueresear, Ecully, France, TJ11.Jessie@gmail.com, Yeming Gong Ljubljana, 1000, Slovenia, mitja.stiglic@ef.uni-lj.si, Mirko Gradisar, Niels Agatz, Martin Savelsbergh

204