2015 Informs Annual Meeting

POSTER SESSION

INFORMS Philadelphia – 2015

37 - Rating Aggregation in Multi-dimensional Rating Systems: How do Reviewers Form Overall Ratings? Dominik Gutt, PhD Candidate, University of Paderborn, Warburger Str. 100, Paderborn, 33098, Germany, dominik.gutt@wiwi.uni-paderborn.de, Dennis Kundisch A recent strain of literature on online product reviews has focused in particular on multi-dimensional product reviews. Multi-dimensional product reviews usually allow the reviewer to rate a product first, based on one overall rating, and second, based on a set of several sub-dimensions. Mostly, overall ratings do not equal e.g. the calculated mean of the sub-dimensions. Our research will shed light on the question, which heuristics reviewers use to form an overall rating. 38 - Cyclic Timetabling and Platforming of Mixed Train Types on a Bidirectional Railway Line We present a mixed integer program for cyclic train timetabling and platforming problem. It is assumed that there are different types of trains moving on a single, bidirectional railway line. Two objectivesótimetable cycle length and total journey time of all train typesóare considered. Constrains include infrastructure characteristics, safety regulations, and operational rules. Heuristics and exact methods are combined with math model to solve large problem instances are solved and discussed. 39 - The Impact of Platform Update Interval on Platform Diffusion in a Cooperative Mobile Ecosystem Yoo S. Hong, Professor, Seoul National University, 1 Gwanak-ro, Gwanak-gu, Seoul, 08826, Korea, Republic of, yhong@snu.ac.kr, Gyesik Oh Periodic platform upgrading is one of the effective strategies for potential participants as well as existing stakeholders. However, the diffusion of an updated platform is ineffective in a cooperative mobile ecosystem where a manufacturer modifies a platform released as open-source software on its smartphones. This paper models the effectiveness of platform diffusion with respect to the platform update interval in consideration of the manufacturer’s decision mechanism. 40 - Operational Health Information Exchange Platform Kang-yu Hsu, PhD Student, Purdue University, 610 Purdue Mall, Health Information Exchange (HIE) has been focusing on utilizing patient electronic health records to improve service quality and patient safety. The exploitation of HIE, however, can go beyond patient records. This study aims to build an infusion drug limit library database and embedded analytical tools as a pilot for the operational HIE platform that may house and utilize other types of data with similar usage and functionality. 41 - Design and Pricing of Crop Insurance Based on NDVI Jing Huang, Hohai University, No.8 Focheng Road, Nanjing, China, huangjingshow@hotmail.com, Anna Shi, Huimin Wang, Jinping Tong The crop index-insurance is a possible approach to solve the dilemmas in traditional crop insurance. In this study, the crop insurance contract is designed based on remote sensed Normalized Difference Vegetation Index (NDVI). Then the crop yield estimation models in different expectations is proposed. Lastly, the pure rate of premium is presented under different triggers using nonparametric kernel density model. The paddy rice in three cities in Yunnan Province, China are chosen as an example. 42 - An Infinite Time Horizon Portfolio Optimization Model with Delays Azmat Hussain, PhD Student, North Carolina State University, Raleigh, NC, 27606, United States of America, ahussai@ncsu.edu We consider a portfolio optimization problem over an infinite time horizon. The problem is formulated as a stochastic control problem and the state is governed by a stochastic process with delay. The goal is to choose investment and consumption controls such that the total expected discounted utility is maximized. Under some conditions, the corresponding HJB equations for exponential, log and HARA utility functions are solved and verification results and the optimal control are also established. 43 - Optimization of the Order-up-to and Reorder Levels for Each Item in the Forward Area of Warehouse We propose an optimization procedure to solve a forward reserve allocation problem and determine the order-up-to and reorder levels for each item in the forward area. In the target warehouse or distribution center, the forward area stores items in a unit of piece for efficient order picking and the reserve area stores items in a unit of case. There is a tradeoff between replenishment frequency and inventory space in the forward area. We show the effectiveness of the proposed method. Takashi Irohara, Professor, Sophia University, 7-1 Kioi-cho, Chiyodaku, Tokyo, 1028554, Japan, irohara@sophia.ac.jp Mojtaba Heydar, University of Newcastle, University Drive, Callaghan, Australia, mojtaba.heydar@newcastle.edu.au West Lafayette, IN, 47907, United States of America, hsu66@purdue.edu, Poching Delaurentis, Yuehwern Yih, Rich Zink

44 - Net Energy Analysis of Wind/Solar Generation and Electric Vehicles: Are They as Green as They Seem? Tongdan Jin, Texas State University, 601 University Drive, San Marcos, TX, United States of America, tj17@txstate.edu, Joseph Duke Wind and solar energy are widely hailed as “free” and “clean” energy. Similarly, electric vehicles (EV) that run on batteries are considered to be much greener than their gasoline counterparts. But is it possible that the environmental benefits from new energy and transport technologies may be exaggerated? The goal of this research is to perform an analysis of the lifecycle CO2 emissions, from wind turbines, solar photovoltaic panels, and EV, to weigh together the cost and benefits. 45 - Design and Modelling of Crowdsource Enabled Relay System for Urban Parcel Delivery Nabin Kafle, Research Assistant, UIC, 842 W Taylor St, Chicago, IL, 60616, United States of America, nkafle3@uic.edu, Bo Zou, Jane Lin We propose a new relay-based intermodal system for urban parcel delivery enabled by crowdsourcing technology. A mathematical model is formulated which (i) identifies relay locations and their visiting order by the delivery truck; and (ii) uses local green modes in place of the delivery truck for last-leg pickup/delivery from the chosen relay points, where crowdsources are selected through auction. We develop a heuristic algorithm to solve the model. 46 - Automatic Sentiment Words Extraction from Online Reviews for Product Comparison Online customer reviews are potentially a great source for improving the quality of products or services. However, it is technically tough to automate it because the review data is vast and unstructured. We propose a text mining approach that can handle a context dependent sentiment of a word in reviews. A word is associated with different sentiment values for different product categories. This enables us to extract pros and cons of a product when compared to similar products. 47 - A Cooperative Game Theory Approach to Highway Cost Allocation Saurav Kumar Dubey, PhD Student Department of Industrial and Systems Engineering, University of Tennessee at Knoxville, 1615 Laurel Avenue, Box-203, Knoxville, TN, 37916, United States of America, skumardu@vols.utk.edu, Alberto Garcia-Diaz Traditional methods of allocating highway cost among vehicle classes include proportional and incremental procedures. Two non-traditional game-theoretic concepts of Least Core and Discrete Aumann-Shapley Value will be shown to be effective in integrating Traffic Capacity and Pavement Thickness requirements. The Nucleolus solution concept will be implemented to strengthen the core stability of Highway Cost Allocation game and reinforce the principles of fairness and rationality. 48 - Incorporating the Impact of Product Reviews on Future Propensity to Purchase Matthew Lanham, Doctoral Candidate, Virginia Tech, Dept. of BIT (0235), Blacksburg, VA, 24061, United States of America, lanham@vt.edu, Ralph Badinelli This study investigates the potential impact that publically accessible product reviews might have on future sales and propensity to purchase for a national retailer and provides a means to incorporate this knowledge posteriori to already derived demand forecasts. The study is motivated by the emergence of Big Data Analytics (BDA) in the retail industry to improve consumer experience, products, and services. 49 - Active Fund Performance: A Full-distributional Analysis Yongjae Lee, Korea Advanced Institution of Science and Technology, 291 Daehak-ro, Yuseong-gu, Daejeon, 34141, Korea, Republic of, yongjae.lee@kaist.ac.kr, William T. Ziemba, Woo Chang Kim Performance evaluation of active funds has been eagerly studied by many scholars in finance. They have compared the performance distribution of active funds to relevant benchmarks. Most of these studies were naÔve comparisons of a full distribution to a small number of points that do not reveal distributional characteristics. Therefore, we analyze the Sharpe ratio distributions of active funds in comparison with analytically derived Sharpe ratio distributions of all feasible portfolios. Eunji Kim, Data Mining Lab., 39-339, Seoul Natl’ Univ., Gwanak-ro, Gwanak-gu, Seoul, Korea, Republic of, eunjikim@dm.snu.ac.kr, Sungzoon Cho

202