Informs Annual Meeting 2017

WC41

INFORMS Houston – 2017

WC39

WC40

352D Multicriteria Decision Making Contributed Session Chair: Jawad Hassan, Kuwait University, Kuwait City, Kuwait, j.hassan@ku.edu.kw 1 - Green New Products Investment Based on Consumer’s Utility Xiaojun Lu, Xi’an Jiaotong University, Xi’an, China, looxiaojun@163.com, Jun Lin This paper explores whether a firm should invest in the new green product to improve their product line based on the direct and spillover utility of green products for consumers, and how the consumers’ utilities affect the green product investment and pricing. The results show that with the increase of consumers’ green attentions, the firm should invest in green product; without spillover effect, consumers’ direct utility will affect the optimal decision. The firm will use low- priced green product to maximize profits when the direct effect is large enough, otherwise with a high-priced green product to maximize profits; when considering spillover effect, the firm will use high-priced green product. 2 - Blockbusters of Video Games: Pricing and Quality Decisions in Two-stage Launch of Triple a Games Gulver Karamemis, University of Rhode Island, 7 Lippitt Rd. 207 Ballentine Hall, Kingston, RI, 02881, United States, gkaramemis@uri.edu, Yuwen Chen, Hee Yoon Kwon In video gaming industry, AAA games are considered to have the highest development costs offering different quality levels in the beta version and launch version of the product. We investigate the developer’s pricing and quality decisions in the two-stage launch of the product. 3 - Decision Optimization on the Cloud Sebastien Lannez, Principal Engineer, FICO, Birmingham, United Kingdom, sebastienlannez@fico.com With FICO Decision Optimizer a strategy analyst can graphically define and optimize decision problems. The interactions between decisions and constrained metrics enables a simple visualization of the assignment of actions to customers, such as investment options or transaction authorizations. The availability of the software within a cloud based environment allows for simple integration into complex IT environments, and offers the ability to exchange and process large amounts of data using sampling and segmentation. 4 - Managing Software Product Innovations Through Versioning Ishwar K.Murthy, Professor, Indian Institute of Management, In this research we examine the issue of introducing a new software product into the market. First, we examine a monopoly situation, wherein the firm considers the option of versioning the new product, simply to learn the market potential. This typically entails an initial period of ‘learning’ followed by value extraction. The question is how to learn about the market potential through versioning. We present a dynamic programming framework to determine the optimal timing of when to introduce which versions. We examine the same problem in the presence of competition. 5 - Excel Based Tool for Interactive Optimization Models Anthony Downward, University of Auckland, Level 3, 70 Symonds Street, Auckland CBD, Auckland, 1024, New Zealand, a.downward@auckland.ac.nz In this talk we will demonstrate a new tool for creating interactive optimization problems in Excel using Javascript. We will show examples of scheduling and routing applications utilizing the SolverStudio modelling framework for Excel. 6 - Integrated Parameter and Tolerance Designs with Manufacturing and Marketing Considerations Jawad Hassan, Assistant Professor, Kuwait University, Kuwait City, Kuwait, j.hassan@ku.edu.kw, M. Jeya Chandra Parameter and tolerance designs are major components of product design and development. Existing models for product design from each of the two fields of manufacturing and marketing lack the required rigor in assessing the other field’s input to their models. Generally, pricing and market dynamics are not included in most of the models found in the manufacturing literature while product quality is overly simplified in many of the models found in the marketing literature. In this study, new integrated parameter and tolerance designs models are proposed that bridge the gap between manufacturing and marketing design considerations and decisions. 103NF Block IIM.Campus, Bannerghatta Road., Bangalore, 560076, India, ishwar@iimb.ernet.in, Giri Kumar Tayi

352E Large-scale Network Modeling and Simulation Sponsored: TSL, Urban Transportation Sponsored Session Chair: Ehsan Jafari, University of Texas, Hart Lane, Austin, TX, 78731, United States, ejafari@utexas.edu 1 - Network Design Problem: A Decentralized Approach Cesar Yahia, University of Texas, Austin, TX, 78731, United States, cesaryahia@utexas.edu, Ehsan Jafari, Stephen Boyles We develop a decentralized algorithm for the transportation network design problem, based on a spatial partition of the network. In the master problem, a regional agent decides the funding allocation between subnetworks, and the subproblems are network design problems for each subnetwork. We develop a solution algorithm based on a sensitivity-analysis heuristic, and test it on two different networks. In addition to computational advantages, the proposed algorithm can be used to model the interactions between different entities with potentially conflicting objectives. 2 - Network Partitioning Algorithms to Reduce Computation Time for Parallel Traffic Assignment Problems Cesar Yahia, University of Texas at Austin, Austin, TX, United States, cesaryahia@utexas.edu, Venktesh Pandey, Ehsan Jafari, Stepehn Boyles Recent methods in the literature to parallelize the traffic assignment problem consider partitioning the network into subnetworks to reduce the computation time for large-scale networks. We propose several network partitioning algorithms which consider the impact of number of subnetworks and the choice of subnetwork boundary on the computation time. Our analysis compares the performance of proposed algorithms on different large-scale networks and makes recommendations on developing appropriate network partitions. 3 - Improving Incidence Responses on Large Network using Simulated Optimization Methods Shanjiang Zhu, George Mason University, Nguyen Engineering Building, Suite 1300, 4400 University Drive Ms 6c1, Fairfax, VA, 22030, United States, szhu3@gmu.edu Traffic incidents are a major cause of urban transportation congestion. Many incident response strategies, including opening of shoulder lanes, rerouting traffic through variable message signs, adjusting ramp metering rate, have been considered in the research and in practice. However, it is a challenge to efficiently identify the best incident response strategies among a large number of strategy combinations. This study combines the simulation-based optimization method and traffic simulation models to reduce the searching space and to efficiently identify the best response strategies. 352F Optimization, Heuristic Programming Contributed Session 1 - Resilient Design and Operation of Offgrid Microgrids with Generation Uncertainty Sreenath Chalil Madathil, Clemson University, 324 Village Walk Ln, Clemson, SC, 29631, United States, schalil@g.clemson.edu, Harsha Nagarajan, Russell Bent, Scott J. Mason, Sandra D. Eksioglu As demand for more renewable energy resources increases in remote communities to reduce carbon footprint, we focus on the design of reliable and cost-efficient offgrid microgrids. We consider the inherent uncertainty in wind and solar power generation. To address these uncertainties, we develop a stochastic model with N-1 security to analyze this challenging problem. We propose two decomposition algorithms and present experimental results using both the benchmark IEEE 13 node network and a real microgrid case study from Alaska. Our results demonstrate the efficacy of our rolling horizon approach. 2 - A Decomposition Method for Convex Optimization Problems the Bienstock Zuckerberg Algorithm Revisited Renaud Pierre Chicoisne, Visiting Professor, University of Colorado denver, 1201 Larimer Street, Academic Building, Office AB1-4122, Denver, CO, 80204, United States, renaud.chicoisne@gmail.com In this talk, we will briefly introduce the Bienstock-Zuckerberg algorithm as it originally appeared for open pit mining scheduling problems. We then discuss its equivalence with a specialized column generation scheme and how its framework can be generalized to solve convex optimization problems. We illustrate this generalization with a resource constrained nonlinear objective routing problem. WC41

517