2016 INFORMS Annual Meeting Program

TB77

INFORMS Nashville – 2016

2 - Modeling Allocation Of Project Resources In Multiproject Portfolio Zinovy Radovilsky, Professor of Management, California State University, East Bay, 25800 Carlos Bee Blvd., Hayward, CA, 94506, United States, zinovy.radovilsky@csueastbay.edu, Vishwanath Hegde We introduced a conceptual model of modeling resource allocations in a multi- project portfolio over its projects lifetime. Using practical resource data in a multi-project setting, we demonstrated that resource allocation patterns can be captured by parametric regression models before and after projects’ due dates.This enables us to accurately forecast resource allocations during projects lifetime. 3 - Resource Allocation And Revenue Management For Age-based Products Hossein Jahandideh, PhD Student, UCLA Anderson School of Management, 3777 Mentone Avenue, Apt 405, Los Angeles, CA, 90034-6473, United States, hs.jahan@gmail.com, Christopher S Tang, Kevin F McCardle, Behnam Fahimnia The value of certain products such as whiskey increases with age. For such products, introducing a new age to the market means introducing a whole new product with demand uncertainty and substitution effects on existing products. Assuming that the firm is able to start the aging process of a set number of barrels every year, we study the question of what fraction of this capacity to allocate to the new age. The goal is to maximize the expected revenue extracted from a fixed yearly production capacity. 4 - Solving Resource Station Location-routing Problem In Emergency Evacuation Through A Resource-space-Time Network Representation Lei Bu, Institute for Multimodal Transportation, Jackson, MS, United States, leibu04168@gmail.com, Zhibin Jiang, Feng Wang, Xing Fu, Chuanzhong Yin Based on a representation of discretized resource-space-time networks, a formulation is proposed to optimize dynamic bus station location and routes decisions in an emergency evacuation of subway station. The proposed integer linear programming formulation could effectively build the modeling representation of time window, resource change and passenger travel distance constraints through a multi-dimensional network with an objective function to minimize the total travel cost. A Lagrangian relaxation approach is utilized to solve the problem. A case study of subway and bus station network in Lianhu District, Xi’an City in China verifies the effectiveness of the model and algorithm. Chair: Lauren Gardner, Senior Lecturer, University of New South Wales, Kensington Campus, Building H20, Sydney, 2052, Australia, l.gardner@unsw.edu.au 1 - Application Of Linear Programming In Dimension Stone Industry Gangaraju Vanteddu, Associate Professor, Southeast Missouri State University, Harrison College of Business, One University Plaza, MS 5815, Cape Girardeau, MO, 63701, United States, gvanteddu@semo.edu A typical dimension stone business unit has to contend with many unique demand and availability related characteristics and constraints, which makes the application of Linear Programming techniques an ideal solution in a wide variety of contexts. In this research, a generic MILP model is proposed to maximize revenue in the presence of demand, operational and technical constraints. 3 - Portfolio Model For Natural Gas Combined Cycle Power Plant Asiye Ozge Dengiz, Research Assistant, Baskent University, Baskent Univesity Industrial Engineering Bagl, Ankara, 06810, Turkey, aokarahanli@baskent.edu.tr, Mehmet Gulsen, Orhan Dengiz Among different types of power generation facilities, natural gas power plants (NGPP) get considerable attention because of the advantages of being able to generate on demand and location flexibility. The producers, often operating several generators, need to make simultaneous planning for their entire plant portfolio to maximize their profit based on the information coming from the market of Turkey and equipment characteristics. For this purpose, in this study for the NGPP, a model is developed for producers to plan generation for a certain horizon considering operation costs and forecasted market data. TB77 Legends E- Omni Opt, Integer Programing II Contributed Session

4 - A Generalized Framework For The Estimation Of Edge Infection Probabilities

Lauren Gardner, Senior Lecturer, University of New South Wales, Kensington Campus, Building H20, Sydney, 2052, Australia, l.gardner@unsw.edu.au, Andras Botas Most network-based infection spreading and diffusion models require a real value or (transmission) probability on the edges of the network as an input, which is often unknown in real-life applications. This work presents a general framework to estimate the value of these probabilities on a network exposed to an infection process, where spatiotemporal information on the outbreak pattern is known. This general model works with a range of infection models, and is able to handle an arbitrary number of observations on such processes.

TB78 Legends F- Omni Opt, Network II Contributed Session

Chair: Parimal Kulkarni, Manager, Supply Chain Analytics, BJC Healthcare, 8300 Eager Rd, Suite 500 D Mailstop 92-92-277, St Louis, MO, 63144, United States, pskf44@umsl.edu 1 - Varying Routes For The Bus Driver’s Sanity Problem Paul Hadavas, Associate Professor, Armstrong State University, 11935 Abercorn Street, Savannah, GA, 31419, United States, Paul.Hadavas@armstrong.edu, Jeremy Dyal The bus driver’s sanity problem is a graph theoretic problem with variable weighted edges. A bus driver needs to minimize the total kid exposure based on kid-minutes. This amounts to summing each kid’s time spent on the bus. The graph characteristics (time to the next stop, number of kids aboard) change once kids are dropped off at a particular stop. In this talk, we expand the possible routes the bus driver can take, including cul-de-sacs and tree-like structures with multiple stops located off the main road, and discuss solution techniques for optimal or near-optimal routes. 2 - A Lagrangian Heuristic For A Rapid Transit Line Design Problem Souhaïla El Filali, University of Montreal, 3520-2920, Chemin de la Tour, Montreal, QC, H3T 1J4, Canada, souhaila.elfilali@cirrelt.ca, Bernard Gendron, Gilbert Laporte We propose a tight formulation for the rapid transit line design problem, which consists of locating stations and segments between them to form a line, with the objective of maximizing O-D pairs coverage under topological and budget constraints. We develop a Lagrangian heuristic to solve the problem, and we test it on artificial and real-life instances. 3 - Search For An Immobile Target On An Undirected Unit Network Songtao Li, Tsinghua University, 519A, Shunde Building, Tsinghua University, Haidian District, Beijing, 100084, China, list14@mails.tsinghua.edu.cn, Simin Huang We consider the problem of searching an immobile target on an undirected network with unit length links. In this problem, multiple searchers traverse the network to find the target. When at least one searcher crosses the target point, the detection happens and the search process end. Our purpose is to optimally guide those searchers to the target ( with known distribution ) to minimize expected search duration. A binary integer programming model is built, and primary computation results is given. Based on the computation results, we discuss the influence of searcher number and targets distribution to the expected search duration. 4 - Supporting Campus Evacuation Decisions Via Network Optimization Jorge Huertas, Graduate Student, Universidad de los Andes, Carrera 1 Este No. 19 A - 40, Bogotá, 111711, Colombia, ja.huertas1845@uniandes.edu.co, Daniel Duque, Ethel Segura, Raha Akhavan-Tabatabaei, Andres L Medaglia In this work we evaluate different emergency scenarios to support an evacuation plan in a campus comprised of multiple buildings. With a time-space network and a Geographical Information System (GIS), we model a campus with 1.7 million square feet that holds up to 27,000 people in a critical moment. To find an evacuation plan, we formulate a MIP based on a minimum cost flow problem formulation with side constraints. To support campus managers, we visualize the solutions under various evacuation scenarios at different scales.

290