2015 Informs Annual Meeting

WB18

INFORMS Philadelphia – 2015

WB19 19-Franklin 9, Marriott Retail Analytics and Optimization Sponsor: Computing Society Sponsored Session

3 - Optimizing Electric Bus Operations and Charging Station Deployment in Singapore Tachun Lin, Assistant Professor, Bradley University, 1501 W Bradley Ave, Peoria, IL, 61625, United States of America, djlin@fsmail.bradley.edu, Zhili Zhou In this study, we build a framework for electric bus deployment in urban area, which supports charging facility deployment and impacts analysis on both traffic network and power grid with limited data sources. This framework can be utilized as a test bed for cities considering electric bus adoption and as a fundamental structure for exploring the impacts of electric vehicles to local power distribution networks. WB18 18-Franklin 8, Marriott Optimization Combinatorial II Contributed Session Chair: Chong Hyun Park, PhD Candidate, Purdue University, 403 W. State St., West Lafayette, IN, 47907, United States of America, park456@purdue.edu 1 - Procedures for The Bin Packing Problem with Precedence Constraints Jordi Pereira, Universidad Adolfo Ibáñez, Avda. Pedro Hurtado 750, Viña del Mar, Chile, jorge.pereira@uai.cl The bin packing problem with precedence constraints is a recently proposed variation of the bin packing problem, which corresponds to a basic model featuring many underlying characteristics of several scheduling and assembly line balancing problems. In this work we propose a dynamic programming based heuristic, and a modified exact enumeration procedure. These methods use several new lower bounds and dominance rules. The results show the effectiveness of the proposed methods. 2 - Does Road Network Density Matter in Optimally Locating Facilities? The aim is to investigate how the density of a road network affects solutions of heuristics by applying the specific case of p-median model in finding optimal location of facilities. The specific experiments are conducted by optimally locating 5 to 50 facilitates on a complex road network of Dalarna, Sweden. Two different heuristics being the vertex-substitution method and the simulated annealing algorithm are applied to solve the p-median problem to have a benchmark and validated performance. 3 - Statistical Bounds in Combinatorial Optimization Xiangli Meng, Dr, Dalarna University, Dalarna University, Falun, Da, 79188, Sweden, xme@du.se, Kenneth Carling We use statistical optimum estimation techniques (SOETs) to assess the quality of heuristic solutions in combinatorial optimization. We examine the performance of different implementations of SOETs and compare with deterministic bounds. Performance is assessed by extensive computer experiments on test problems. We find SOET to give (substantially ) tighter gap that deterministic bounds, but SOET needs to be applied cautiously. 4 - Parametric Approaches to Fractional Combinatorial Problems: Analytical and Computational Studies Chong Hyun Park, PhD Candidate, Purdue University, 403 W. State St., West Lafayette, IN, 47907, United States of America, park456@purdue.edu, Yanjun Li, Robert Plante A parametric modeling approach provides effective technique for obtaining optimal solutions of the linear fractional combinatorial optimization problems. We consider two algorithms for solving the parametric model and investigate the efficiency of the algorithms both theoretically and computationally. For the computational study, the algorithms are used to solve fractional knapsack problems and are compared to other algorithms (e.g., Newton’s method). 5 - Combinatorial Auctions with Items Arranged in Rows Dries Goossens, Ghent University, Tweekerkenstraat 2, Gent, 9000, Belgium, Dries.Goossens@ugent.be, Bart Vangerven, Frits Spieksma We consider combinatorial auctions of similar goods (seats, land, ...) that can be arranged in rows. We describe a dynamic programming algorithm which, for a 2- row problem with connected and gap-free bids, solves the winner determination problem optimally in polynomial time. We also study a number of extensions, and generalize our result to a setting with connected bids in a 3-row problem. Finally, we study the complexity for bids in a grid, complementing known results in literature. Johan Hakansson, Professor, Dalarna University, Sweden, Hügskolan Dalarna, 79188 Falun, Falun, 79188, Sweden, jhk@du.se, Pascal Rebreyend, Xiaoyun Zhao

Chair: Tulay Flamand, University of Massachusetts, Amherst, Isenberg School of Management, 121 Presidents Drive, Amherst, MA, 01003, United States of America, tulayvarol@gmail.com 1 - Maximizing Impulse Buying via Store-wide Shelf Space Analytics Bacel Maddah, Associate Professor, American University of Beirut, Beirut, Beirut, Lebanon, bacel.maddah@aub.edu.lb, Tulay Flamand, Ahmed Ghoniem Impulse (unplanned) buying constitutes a common shopping behavior. We investigate how retailers can optimize product shelf allocation in a fashion that improves product visibility to consumers and maximizes impulse buying. We examine the interplay between a retail store layout, the location of products, and their allocated shelf space with the notion of impulse buying. Specifically, we develop and analyze a mixed-integer nonlinear program (NLP) that allocates shelf space to product categories. 2 - Optimization Approaches for Generalized Assignment Problems with Location/allocation Considerations Tulay Flamand, University of Massachusetts, Amherst, Isenberg School of Management, 121 Presidents Drive, Amherst, MA, 01003, United States of America, tulayvarol@gmail.com, Ahmed Ghoniem, Mohamed Haouari We address a novel type of generalized assignment problems with location/allocation considerations that arise in retail shelf space allocation. Single- and multiple-knapsack variants of this problem are formulated along with modeling enhancements. Our proposed branch-and-price algorithm yields significant computational savings over the branch-and-bound/cut algorithm in CPLEX for challenging instances. 3 - Dynamic Assortment Planning under Cross-selling and Cannibalization Effects Ameera Ibrahim, Assistant Professor, Saint Mary’s College of California, 1928 St. Marys Rd, Moraga, CA, 94556, United States of America, ai7@stmarys-ca.edu, Ahmed Ghoniem, Bacel Maddah We study the problem where a decision-maker optimizes the assortment and release times of products that belong to different categories over a multi-period horizon. Products have a longevity over which their attractiveness decays while being positively or negatively impacted by the specific mix of products that were introduced. We propose a 0-1 fractional program that employs an attraction demand model. A mixed-integer linear reformulation is developed that enables exact solutions to the problem. WB21 21-Franklin 11, Marriott Operations Research Applications in Vaccine Pricing and Distribution Sponsor: Health Applications Sponsored Session Chair: Maryam Hasanzadeh Mofrad, University of Pittsburgh, 1048 Benedum Hall, Pittsburgh, 15261, United States of America, hasanzadeh.mofrad@gmail.com 1 - Exploring Market Segmentation in a Centralized Vaccine Market under Stochastic Reservation Prices Galo Mosquera, Vaccine Access And Affordability In a Centralized Market Under Stochastic Reservation Prices, Rochester Institute of Technology, 81 Lomb Memorial Drive, Rochester, NY, 14623, United States of America, gem9454@mail.rit.edu, Ruben Proano We consider a vaccine market in which, a monopsonistic entity aims to maximize total social surplus and the willingness to pay for different vaccines are stochastic. Preliminary experimental results show that increasing the number of market segments has undesirable effects on the profitability and affordability of key market segments.

406