2015 Informs Annual Meeting

POSTER SESSION

INFORMS Philadelphia – 2015

50 - Behavioral Ordering Decision under Downward Substitution Yan Li, Dr., China University of Mining and Technology, Ding 11, Xueyuan Road, Beijing, China, liyan@cumtb.edu.cn, Bojiao Mu Downward substitution is one common strategy for selling multi-class products. The previous research assumes perfect rationality. This paper relaxes the assumption and utilizes the MNL model to depict the ordering behaviors for substitutable products. We compare the ordering quantity considering substitution with that without substitution. The substitution effect shows non- monotonicity regarding the extent of rationality and is superior to the one predicted by rationality. 51 - Optimal Capital Structure and Credit Spread under Partial Information Bo Liu, UESTC, No.4, Section 2, North Jianshe Rd., Chengdu, China, liub@uestc.edu.cn The paper first incorporates partial information friction to extend the classic optimal capital structure model. We derive closed-form results for the value of risky debt,credit spread, default threshold, and for optimal capital structure.We find that under partial information,dynamic learning significantly increases the optimal coupon level and firm’s leverage, and improves the tax advantage to debt. 52 - Optimal Production and Inventory Policy in Solar Photovoltaic Supply Chain Xiangrong Liu, Bridgewater State University, 95 Grove Street, Bridgewater, PA, United States of America, Xiangrong.Liu@bridgew.edu, Chuanghui Xiong The development and utilization of solar photovoltaic (PV) energy has progressed at a very fast pace. With decreasing price of PV module and uncertain government incentives, this research models the production and inventory strategies in the setting of a PV supply chain with a PV manufacturer, an installer and an end customer. Based on the manufacturer’s and installer’s optimal decision, this study discusses how to improve supply chain performance through parameters setting in contract design. 53 - Optimal Stopping Game with Investment Spillover Effect Akira Maeda, Professor, The University of Tokyo, 3-8-1 Komaba, Meguro, Tokyo, 153-8902, Japan, maeda@global.c.u-tokyo.ac.jp, Motoh Tsujimura, Ryuta Takashima The purpose of this study is to analyze the game over optimal choice of firm’s investment time, focusing on the case that there is positive externality in the effect of investment. We consider a situation where firms can increase their subsequent revenue stream by making an irreversible investment, and the investment has a spillover effect to other firms. This setup describes gaming over optimal stopping problems. We examine the property of the subgame perfect Nash equilibrium. 54 - Recent Trends in Blood Banking Systems: A Supply Chain Perspective Amir Masoumi, Assistant Professor Of Management, Manhattan College, 4513 Manhattan College Parkway, DLS 504, Riverdale, NY, 10471, United States of America, amir.masoumi@manhattan.edu Blood service operations are a key component of the healthcare system all over the world. In the US prior to 2008, there were several reported cases of blood shortages; however, the scenario has significantly changed thereafter. The total number of whole blood and red blood cells collected annually decreased from 17.3 to 15.7 million units during the 2008-2011 period. We investigate the recent trends in supply and demand management of blood banking systems from a logistics perspective. 55 - Optimal Sizing of a Price-maker Energy Storage Facility Considering Uncertainty Ehsan Nasrolahpour, University of Calgary, 2500 University Dr. NW, Calgary, AB, Canada, enasrola@ucalgary.ca This paper proposes a strategic investment model for a price-maker energy storage facility considering market uncertainties. The proposed model is a stochastic bi- level optimization problem where planning and operation decisions of the energy storage facility are made in the upper level, and market clearing is modeled in the lower level under different operating conditions. The bi-level optimization problem is recast as an Mathematical Program with Equilibrium Constraints (MPEC). 56 - Army Materiel Systems Analysis Activity (AMSAA) Joseph Olah, AMSAA, 392 Hopkins Road, APG, MD, United States of America, joseph.m.olah.civ@mail.mil, Tiffany Gutowski AMSAA is the Army’s independent source of data, modeling & simulation, and materiel lifecycle & logistics systems analysis to support the Army’s Equipping, Sustaining and Warfighting decisions. AMSAA’s Core Competencies are Independent Materiel Performance and Effectiveness Analysis, Independent Logistics Analysis, Field Data Collection and Analysis, Program Management of DoD’s JTCG-ME Program, Strategic/Corporate Level Decision Analysis, and Certified System Level Performance Data.

57 - Reinforcement Learning Algorithm for Blood Glucose Control in Diabetic Patients Mahsa Oroojeni Mohammad Ja, Northeastern University, 334 Snell Engineering, Boston, MA, United States of America, oroojeni.m@husky.neu.edu In this paper a reinforcement learning algorithm is proposed for regulating the blood glucose level of Type I diabetic patients. In the proposed reinforcement learning algorithm body weight and A1C level define the state of a diabetic patient. For the agent, insulin dose levels constitute the actions. As a result of a patient’s treatment, after each time step t, the patient receives a numerical reward depending on the response of the patient’s health condition. 58 - Modeling the Stockist Omkar Palsule Desai, Associate Professor, Indian Institute of Management Indore, Prabandh Shikhar, Rau Pithampur Road, Indore, MA, 453556, India, omkardpd@iimahd.ernet.in, Ananth Iyer We focus on the problem of distribution to the millions of small shops that constitute the retail sector in India, as well as many other developing countries. We model the role of a stockist - a supply chain entity whose role is to facilitate distribution. We use a principal agent model structure, with a complements or substitutes relationship between manufacturer assistance and retailer impact, to understand the optimal contract structure, i.e., level of assistance and associated retail margin. 59 - Automatic Design of Methods for Combinatorial Optimization Problems Lucas Parada, General Manager, Universidad de Concepcion, Avenida Inglesa 134 / 504, Concepcion, 4040409, Chile, lucasparada20@gmail.com Designing an method to solve an optimization problem is a complex intellectual task. However, to design an algorithm is also an optimization problem. To solve this second level problem we combine and evolve elementary algorithmic components through genetic programing. The produced algorithms show promising features such as low solution errors and small computational times for several classical optimization problems. 60 - Bayesian Adjusted Uplift Modeling for Direct Mail Campaign Yidong Peng, Conclusive Analytics, 13620 Reese Boulevard E. Suite 300, Huntersville, NC, 28078, United States of America, yidong.peng@ndsu.edu The study compares the performance of traditional respond model, uplift model and our proposed Bayesian adjusted uplift model on selecting customers for direct mail campaign. The proposed model applies customers’ responses to historical campaign to generate the posterior uplift estimates based on result of uplift model. A case study is conducted to verify that the proposed model provides higher sales lift by using the real monthly directly campaign data from a top auto- parts retail company. 61 - How to Make Big Blue (IBM) Business Segments Fast and Responsive Alan Piciacchio, Senior Technical Staff Member / Lead Request For Service Business Analyst In Rfs, IBM, 2455 South Road, Poughkeepsie, NY, 12590, United States of America, alanpic@us.ibm.com, Jose Cano, Skip Jahn This poster will describe how a big company like IBM can be nimble and fast and responsive. Over the past 3 years - in the growth segment (hundreds of millions of dollars yearly) of IBM’s Global Technology Services unit, an impactful set of analytics and actions have been deployed to dramatically improve business revenue by tens of millions of dollars, via a 65% improvement in cycle time. 62 - Continuum Approximation Modeling of Freight Distribution Systems Mahour Rahimi, University of Massachusetts, Amherst, 139 Marston Hall, 130 Natural Resources Rd., Amherst, MA, 01003, United States of America, mrahimi@umass.edu, Eric Gonzales This study presents a continuous approximation model for truck deliveries which relate the operating parameters to the characteristics of the service and network, service area, and demand rate. The objective of this study is to minimize the total cost of distributing multicommodity freight from an origin to randomly distributed points, with or without transshipments, and within a limited amount of time. Two different distribution methods are considered: peddling, and peddling with transshipment. 63 - Modeling Relation Between Natural Problems and Formal Structures: A Health Systems Application Edmond Ramly, University of Wisconsin-Madison, 20 Sherman Terrace, Unit 6, Madison, WI, 53704, United States of America, edmond.ramly@gmail.com We formulate a class of cyber-social systems where formal (mathematical) and natural (problem structuring) operations research are complementary and insufficient separately. We adapt the Hertz-Rosen Modeling Relation from systems biology as a unifying framework relating natural and formal systems with encoding and decoding operations. We present a category-theoretic axiomatization and a demonstration of complementarity in a health IT evaluation case.

203