2015 Informs Annual Meeting

WC24

INFORMS Philadelphia – 2015

WC24 24-Room 401, Marriott Constraint and Mixed Integer Programming Sponsor: Artificial Intelligence Sponsored Session Chair: Louis-Martin Rousseau, CIRRELT Ecole Polytechnique de Montréal, CP 6079 Succ Centre-Ville, Montréal, H3C3A7, Canada, louis-martin.rousseau@polymtl.ca 1 - A Constraint Programming Approach for Solving Convex Quadratically Constrained Problems Chris Beck, University of Toronto, 5 King’s College Rd, University of Toronto, Toronto, On, M5S3G8, Canada, jcb@mie.utoronto.ca, Wen-Yang Ku Inspired by the geometric reasoning exploited in discrete ellipsoid-based search (DEBS) from the communications literature, we develop a constraint programming (CP) approach to solve problems with convex quadratic constraints. Such constraints appear in numerous applications. We strengthen the key aspects of DEBS and implement them as combination of a global constraint and variable/value ordering heuristics. Preliminary experiments on a variety of benchmark instances show promising results. 2 - Strengthening Convex Relaxations with Bound Tightening for Power Network Optimization Convexification is a fundamental technique in (mixed-integer) nonlinear optimization and many convex relaxations are parametrized by variable bounds, i.e., the tighter the bounds, the stronger the relaxations. This paper studies how bound tightening can improve convex relaxations for power network optimization. It adapts traditional constraint-programming concepts to a relaxation framework and shows how bound tightening can dramatically improve power network optimization. 3 - Improved Constraint Propagation via Lagrangian Decomposition Andre Augusto Cire, Assistant Professor, University of Toronto Scarborough, 1265 Military Trail, Toronto, ON, M1C 1A4, Canada, acire@utsc.utoronto.ca, David Bergman, Willem-jan Van Hoeve Constraint propagation is inherently restricted to the local information that is available to each propagator. We propose to improve the communication between constraints by encoding them as decision diagrams and applying a Lagrangian Decomposition scheme, where penalty costs are incurred when the variable assignments in each constraint do not correspond to one another. We show that propagating Lagrangian cost information can help improve the quality of bounds as well as the solution time. 4 - General Bounding Mechanism for Constraint Programs Louis-Martin Rousseau, CIRRELT Ecole Polytechnique de Montréal, CP 6079 Succ Centre-Ville, Montréal, H3C3A7, Canada, louis-martin.rousseau@polymtl.ca, Claude-guy Quimper, Hoang Minh Ha This paper presents an approach based on ideas from Lagrangian decomposition (LD) that establishes a general bounding scheme for any CP. We provide an implementation for optimization problems that can be formulated with knapsack and regular constraints, and we give comparisons with pure CP approaches. WC26 26-Room 403, Marriott Project Management I Contributed Session Chair: Ted Klastorin, Professor, University of Washington, Foster School of Business, Box 353226, Seattle, WA, 98195-3226, United States of America, tedk@u.washington.edu 1 - Project Management for IT-driven Development of Disadvantaged Communities in the Developing World Devendra Potnis, Assistant Professor, University of Tennessee at Knoxville, 1345 Circle Park Drive, Communications Bldg., Suite 451, Knoxville, TN, 37996, United States of America, dpotnis@utk.edu Inability of researchers to collect rich data from disadvantaged communities leads to the failure of the projects aiming to develop information systems for the development of disadvantaged communities. This study demonstrates the specific ways in which researchers can contextualize project management principles to address data collection barriers by managing scope, time, cost, human resources, quality, communications, and risks related to “IT for development projects” in the developing world. Pascal Van Hentenryck, University of Michigan, 1205 Beal Avenue, Ann Arbor, MI, 48109, United States of America, pvanhent@umich.edu, Carleton Coffrin, Hassan Hijazi

2 - On Coordinating Contracts in Decentralized Stochastic Projects Ted Klastorin, Professor, University of Washington, Foster School of Business, Box 353226, Seattle, WA, 98195-3226, United States of America, tedk@u.washington.edu, Tony Chen, Michael Wagner We study a decentralized project where a client organization contracts individual tasks to independent subcontractors. We show that a simple linear incentive contract coordinates the project while maximizing the expected project profit and minimizing two risk measures. We show how this contract can be implemented in both bidding and auction environments. 3 - Stochastic Resource-constrained Project Scheduling using Approximate Dynamic Programming Haitao Li, Associate Professor, University of Missouri - St. Louis, 229 ESH, St. Louis, MO, 63121, United States of America, lihait@umsl.edu, Keith Womer We present an effective and efficient approximate dynamic programming (ADP) algorithm for the well-known stochastic resource-constrained project scheduling problems. Our approach features a hybrid ADP framework that integrates both the rollout look-ahead policy and the look-back policy via lookup table to enhance solution performance. 4 - Resource Allocation Patterns in a Multi-project Portfolio of Engineering Projects Vishwanath Hegde, California State University East Bay, 25800 Carlos Bee Blvd, Hayward, CA, 94542, United States of America, vish.hegde@csueastbay.edu, Zinovy Radovilsky Using historic resource loading data in a multi-project setting, we show that resource distribution patterns can be captured by parametric regression models, which can forecast resource distribution during project lifetime using project due date and attributes. Forecast based on historical data instead of priory assumption can improve the accuracy of resource planning. This macro estimation approach can be deployed easily by project portfolio managers. Chair: Sune Lauth Gadegaard, PhD fellow, Department of Economics and Business, Aarhus University, Fuglesangs AllÈ 4, bygning 2622, lokale 315a, Aarhus V, 8210, Denmark, sgadegaard@econ.au.dk 1 - A Multi-Objective Decision Making Model based on Uncertain Preference Sequence Information Xiao Liu, Mr., Huazhong University of Science and Technology, No. 1037, Road Luoyu, Wuhan, 430074, China, liu_xiao@hust.edu.cn, Huimin Ma In this paper, we put forward a multi-objective decision making model for two- sided matching based on uncertain preference sequence information. We first design a mechanism to get preference ordinal value in uncertain sequence. Then multiple optimal objectives on two-sided matching make up a mixed 0-1 integer linear programming model. At last comparing with the typical two-sided matching optimization approaches, we discuss their characteristics and performance in different evolution criteria. 2 - Preferences Modeling in Goal Programming Model through Satisfaction Functions Belaid Aouni, Associate Dean, Qatar University, College of Business and Economics, P.O. Box 2713, Doha, Qatar, belaid.aouni@qu.edu.qa The concept of Satisfaction Functions was developed by Martel and Aouni in 1991 to explicitly integrate the Decision-Maker’s preferences within the Goal Programming Model. Through this concept several criticism of the Goal Programming model were addressed. This concept has been widely applied to different Goal Programming variants and utilized in solving several application cases. The aim of this paper is to present an analysis of different applications of this concept with a guideline. 3 - Balancing Faculty and Student Preferences in the Assignment of Students to Groups Richard Forrester, Associate Professor Of Mathematics, Dickinson College, College and Louther Streets, Carlisle, PA, 17013, United States of America, forrestr@dickinson.edu, Kevin Hutson We develop a practical approach for assigning students to groups where the goal is to create diversity within the groups while considering the preferences of the students. A motivating application is the assignment of students to seminars. Faculty want seminars that are balanced with regards to gender and ethnicity, while students want to be assigned to one of their higher ranked seminars. Our model is a multi-objective convex quadratic program that can be solved using a standard solver. WC27 27-Room 404, Marriott Multicriteria Decision Making I Contributed Session

434