2015 Informs Annual Meeting

SD70

INFORMS Philadelphia – 2015

2 - Optimizing Storage-class Formation in Unit-load Warehouses Yun Fong Lim, Associate Professor, Singapore Management University, 50 Stamford Road, #04-01, Singapore, 178899, Singapore, yflim@smu.edu.sg, Marcus Ang We propose a new approach to optimize storage classes for a unit-load warehouse with a general layout. Under this approach, the “attractiveness” of each storage location is determined by its frequency of visits, which is estimated by a linear program that considers the warehouse’s layout and the products’ arrivals and demands. We group the locations with similar visit frequencies in the same class. Our approach gives a lower average travel cost than a cost-based method and a grid-based method. 3 - Two Single Instruction Multiple Data Implementations for Solving the Quadratic Assignment Problem Clara Novoa, Associate Professor, Texas State University, 601 University Dr., San Marcos, TX, 78666, United States of America, cn17@txstate.edu, Apan Qasem, Abhilash Chaparala We solve the Quadratic Assignment Problem by implementing 2-opt and tabu search in the Graphical Processing Unit (GPU). For the 2-opt we fine tune the thread block configuration and exploit inter-thread data locality through shared memory allocation. In the tabu search we exploit dynamic parallelism. We experiment with QAPLIB data sets. Tabu search accuracy is very satisfactory while 2-opt performance is impressive. Results are contrasted to a Tabu search GPU implementation from other authors. 4 - Optimizing Vehicle Travel Speed in Green Vehicle Routing Problems Xiaoren Duan, Teaching Assistant, University of Louisville, Department of Industrial Engineering, University of Louisville, A Green Vehicle Routing problem with various travel speed is formulated to minimize total carbon emission. Heuristic algorithm based on Savings Algorithm and Tabu Search is developed to solve this problem. Numerical experiments show that the heuristic performs better compared with GAMS and can achieve 15.69% and 32.27% carbon emission reduction compared with basic G-VRP with and without time window limitation respectively. Impact of congestion on carbon emission is also investigated. Louisville, KY, 40292, United States of America, duanxiaoren@gmail.com, Sunderesh Heragu

We develop a probabilistic flow-based location model to optimally deploy electric vehicle (EV) charging stations on traffic network, taking into account the probability of a demand node becoming an EV adopter. We demonstrate the model with the Sioux Falls network and solve the model using a Lagrangian relaxation based algorithm. 2 - Optimization-based Planning of Capacitated Infrastructure for Intercity Trips of Electric Vehicles Yu Nie, Northwestern University, y-nie@northwestern.edu, Ali Zockaie, Mehrnaz Ghamami The main purpose of this study is to facilitate the long-distance trips of electric vehicles. The objective is to minimize the construction cost of charging stations, battery cost, and refueling delay, while maintaining a certain level of service. To this end, a nonlinear optimization model is developed. To overcome computational difficulties of the commercial solvers, a metaheuristic algorithm is proposed to solve the nonlinear model, as the size of problem grows in real world case studies. 3 - Multi-period Capacitated Flow Refueling Location Problem Anpeng Zhang, University at Buffalo, SUNY, 339 Bell Hall, Buffalo, NY, 14228, United States of America, anpengzh@buffalo.edu, Jee Eun Kang, Changhyun Kwon We formulate a new flow refueling location problem for electric vehicles, considering the capacity of rechargers and the time span of construction. The model will help us determine the optimal locations of recharging stations as well as the number of recharging modules at each station over multiple time periods. We develop heuristic methods and present computational experiments based on the freeway network that spans between Washington DC and Boston. 4 - Infrastructure Planning for Fast Charging Stations in a Competitive Market Zhaomiao Guo, University of California, Davis, 614 Sycamore Lane. Apt. 232, Davis, CA, United States of America, zmguo@ucdavis.edu, Yueyue Fan, Julio Deride We study the fast charging infrastructure planning under competition using Multi-agents Optimization Problem with Equilibrium Constraints modeling framework. We find that the investment pattern could be affected by consumers’ weights on charging price and charging availability: if consumers care more about charging availability, the investment may cluster to a few locations; on the contrary, the investment may diffuse through out the network. Chair: Jianjun Shi, Georgia Institute of Technology, 765 Ferst Dr, Atlanta, United States of America, jianjun.shi@isye.gatech.edu 1 - Progressive Measurement and Monitoring for Multi-resolution Data in Surface Manufacturing Considering Cross Correlations Hui Wang, Assistant Professor, Florida State University, 2525 Pottsdamer St, Tallahassee, FL, 32310, United States of America, hwang10@fsu.edu This paper develops a new approach to modeling and monitoring surface variations by fusing in-plant multi-resolution measurements and process information. The fusion is achieved by considering cross correlations among measured data and manufacturing process variables based on cutting dynamics. The model can make Bayesian inference on surface shapes progressively. A new monitoring scheme is then proposed for jointly detecting and locating defects without significantly increasing false alarms. 2 - Prediction of the Failure Interval with Maximum Power Based on the Remaining Useful Life Distribution Junbo Son, PhD Candidate, University of Wisconsin-Madison, 1513 University Avenue, Madison, WI, 53706, United States of America, json5@wisc.edu, Qiang Zhou, Shiyu Zhou, Mutasim Salman Prognosis of remaining useful life (RUL) of a unit or a system plays an important role in system reliability. One key aspect of the RUL prognosis is constructing the best prediction interval. In this paper, we propose a new method, namely maximum prediction power interval (MPI). The MPI guarantees the best prediction performance under the given acceptable error range. A numerical simulation study and case study with real data confirm the better features of MPI over existing prediction intervals. SD72 72-Room 203A, CC IIE Transactions Sponsor: Quality, Statistics and Reliability Sponsored Session

SD70 70-Room 202A, CC RAS Roundtable: Part II Railroad Operations

Efficiency and Recovery Sponsor: Railway Applications Sponsored Session

Chair: Erick Wikum, Principal Scientist, Tata Consultancy Services, 1000 Summit Drive, Milford, OH, 45150, United States of America, erick.wikum@tcs.com 1 - Railroad Operations Efficiency and Recovery Erick Wikum, Principal Scientist, Tata Consultancy Services, 1000 Summit Drive, Milford, OH, 45150, United States of America, erick.wikum@tcs.com, Siddhartha Sengupta, Tao Tang, Lonny Hurwitz In the railroad industry, achieving efficient operations and developing the capability to recover from inevitable disruptions are key to both customer service and financial performance. In this session, the second of two, panelists from the railroad industry worldwide explore how to define and measure efficiency and recovery and share case studies and a vision for the role OR/MS and analytics has played and can play in operational efficiency and recovery.

SD71 71-Room 202B, CC Alternative Fuel Vehicles and Sustainable Transportation I Sponsor: TSL/Urban Transportation Sponsored Session

Chair: Changhyun Kwon, Associate Professor, University of South Florida, 4202 East Fowler Avenue, ENB 118, Tampa, FL, 33620, United States of America, chkwon@usf.edu 1 - A Probabilistic Location Model for Deployment of Electric Vehicle Charging Stations Eric Huang, Assistant Professor, Clemson University, 314 Lowry hall, Clemson, SC, 29634, United States of America, yxhuang@clemson.edu, Shengyin Li

142