2015 Informs Annual Meeting

POSTER SESSION

INFORMS Philadelphia – 2015

5 - Minimizing Automotive Supply Chain Costs under Mixed Transportation Modes Sherif Masoud, Operations Research Analyst, RockTenn, 3950 Shackleford Rd., Duluth, MN, United States of America, smasoudphd@gmail.com, Scott Mason We present an industry-motivated integrated production and transportation problem focused on short-term automotive supply chain planning. We consider multiple, heterogeneous modes of transportation that offer a cost vs. delivery time option to the manufacturer. Computational results demonstrate the efficiency of a proposed metaheuristic-based solution approach, given the problem’s NP-hard computational complexity. Software Demonstration Cluster: Software Demonstrations Invited Session 1 - AMPL - Developing Optimization Applications Quickly and Reliably with Algebraic Modeling Robert Fourer, President, AMPL Optimization Inc., 2521 Asbury Ave, Evanston, IL, 60201, United States of America, 4er@ampl.com Can you negotiate the complexities of the optimization modeling lifecycle, and deliver a working application before the problem owner loses interest? Algebraic languages were invented to streamline the key steps of model formulation, testing, and revision. Today they are supported by powerful facilities for embedding models into larger systems and deploying them to users. This presentation introduces algebraic modeling for optimization through examples using classic and recently introduced features of the AMPL language and system. 2 - Gurobi Optimization, Inc. – Modeling with the Gurobi Python Interface Renan Garcia, Optimization Support Engineer, GAMS Development Corp Are you looking for an environment that combines the expressiveness of a modeling language with the power and flexibility of a programming language? The Gurobi Python interface allows you to build concise and efficient optimization models using high-level modeling constructs. Moreover, Python itself has a vast ecosystem of packages designed to increase your productivity, such as a notebook-style interface (iPython Notebook), data access capabilities and web development tools. This tutorial will provide an overview of these features, including detailed examples that show how to use the Python interface to build models that can be turned into full optimization applications. Monday, 12:30pm - 2:30pm Exhibit Hall A Monday Poster Session Contributed Session Chair: Wenjing Shen, Drexel University, Philadelphia, PA, United States of America, ws84@drexel.edu Co-Chair: Allen Holder, Rose-Hulman Mathematics, Terre Haute, IN, United States of America, holder@rose-hulman.edu Co-Chair: Min Wang, Drexel University, 3141 Chestnut Street, Philadelphia, PA, United States of America, mw638@drexel.edu 1 - Big Data Marwah Halwani, University of North Texas, 2812 Loon Lake Road, Denton, TX, 76210, United States of America, marwahhalwani@my.unt.edu, Victor Prybutok, Adam Corwin, Daniel Peak The Big Data Model developed to provide a foundation for the proper use of Big Data and Data Visualization in Social Media environments that will drive positive bottom-line results. This research addresses how can Big Data represented with Data Visualization in a Social Media environment contribute to better decisions 2 - Brightness-location Congruency Effects on Consumer Behavior in Retail Context Tsutomu Sunaga, Professor, Kwansei Gakuin University, 1-1-155, Uegahara, Nishinomiya, Hyogo, 662-8501, Japan, sunaga@kwansei.ac.jp, Jaewoo Park The study investigates the effects of the crossmodal correspondence between colour and visual heaviness on consumer purchase behavior. The results of the experiments demonstrate that brightness-location congruence, specifically products with bright (dark) colour at the higher (lower) shelf positions, increases shoppers’ perceptual fluency and promotes their purchase behaviour. MB79 79-Room 302, CC

3 - A Kalman Filter Algorithm for Artillery Firing Shift Michael Bendersky, Ben Gurion University of the Negev, Beersheba, Israel, michael.bendersky@gmail.com, Israel David We propose an innovative algorithm for artillery firing shift using the Kalman Filter approach. Firing shift implies an immediate artillery engagement of a target (“fire for effect”) without a prior fire adjustment (by a forward observer). The capability of firing shift provides undeniable operational advantages. Implementing the Kalman Filter allows sequential fire adjustment relying on multiple auxiliary targets. 4 - Demand Forecasting and Area Marketing for Gas Appliances Kosuke Shaku, Tokyogas, 1-5-20,Kaigan, Minato-ku, Tokyo, Japan, shaku@tokyo-gas.co.jp Tokyo gas has been utilizing O.R. to a lot of fields such as marketing, emergency response, and so on. Demand Forecasting of gas appliances has been a big problem for gas appliances sales. We managed to establish the method of quantitatively rational demand forecast by utilizing the CRM data of appliances stocks, survival analysis, and transience of appliances types in replacement. The result of this work has been adopted to many Tokyo gas measures such as sales goal setting and area marketing. 5 - Who to Call Predictive Modeling of Potential Customers Based on Customer Behavior Data Lin Shi, Kihihi Network Information Systems Technology, China, shilin@baixing.com A telemarketing campaign is operated in a classified advertisements website Baixing.com. A model is needed to predict the possibility of customers order and then those with highest possibility can be selected. The paper designed and implemented a practical decision support system which could generate and distribute customer lists to sales representative. The poster introduces the object, business background, specific aims, modeling procedures, modeling data, final result and the conclusions of the work. Four techniques have been used to compare the performance. Random forest gives the best result. The final dynamic model integrates customers online behavior data, which is also called click-stream data as indicators of willingness to pay. The result from field study inspired that in the future work, we may push dynamic modeling for more robust and precise prediction. 6 - The Opportunity Cost of Federal Subsidies for Electricity Generation in the U.S. James Gibson, USMC, 2414 Turtle Bay Dr, New Bern, NC, 28562, United States of America, gibson.james.r@gmail.com This study is the first investigation of the opportunity cost associated with electric utility sector federal subsidies using the mean-variance portfolio theory. The application of portfolio theory provides for an examination of how policy decisions influence electricity generation costs. The results indicate federal subsidies have an uncertain effect on electricity generation costs and the associated tax burden becomes an opportunity cost assumed by society and individual taxpayers. 7 - The Effect of Shape and Semantic Novelty in Product Design Usage Harris Kyriakou, Stevens Institute of Technology, 1 Castle Point on Hudson, Hoboken, NJ, 07087, United States of America, ckyriako@stevens.edu Our study examines 35,727 product designs submitted to the largest 3D printing online community from January 2009 to June 2013, showing how (i) shape novelty, (ii) semantic novelty and (iii) their interaction affect product design usage. We develop both a shape-based and a text-based measure of novelty by identifying designs that are dissimilar to any preexisting design. 8 - Who Wants My Product? Affinity-based Marketing Leyla Zhuhadar, Assistant Professor, Western Kentucky University, 1906 College Heights Blvd, Grise Hall, Bowling Green, KY, 42101, United States of America, leyla.zhuhadar@wku.edu In this research, we have developed a marketing application for data mining with the goal of publicizing a new product to those customers with a high affinity for it. We assembled suitable data to test and evaluate different mining algorithms on it. We have used the buyers of our product of interest as “model customers” for finding similar customers among the non-buyers. Finally, we deployed our final model to our customer base. 9 - Robust Security-constraint Unit Commitment with Dynamic Rating Anna Danandeh, University of South Florida, Tampa, FL, 33613, United States of America, annadanandeh@mail.usf.edu, Bo Zeng, Brian Buckley A challenge in UC is the impact of uncertain factors such as ambient temperature on generation and transmission capacities. Since system capacity is mostly determined statically, weather changes can cause outages and/or congestions. We developed a 2 stage robust security-constraint UC formulation which dynamically rates the assets and hedges against possible efficiency drops. Leveraging the correlation between weather and load, it yields a less conservative decision and a faster computation.

199