2016 INFORMS Annual Meeting Program

MD12

INFORMS Nashville – 2016

MD13 104C-MCC Project and Resource Planning Sponsored: Optimization, Computational Optimization and Software Sponsored Session Chair: Haitao Li, Univ. of Missouri - St Louis, St Louis, MO, United States, lihait@umsl.edu Co-Chair: Cipriano A. Santos, Distinguished Technologist, HPe-IT Center of Excellence, Virtual Office, Palo Alto, CA, 94304, United States, cipriano.santos@hpe.com 1 - An Optimization Approach To Workforce Planning In Professional Service Firms Vincent Hargaden, Assistant Professor, University College Dublin, 209 Engineering & Materials Science Centre, Belfield, Dublin 4, Ireland, vincent.hargaden@ucd.ie, Jennifer K. Ryan, Amir Azaron We develop a comprehensive mixed integer programming model for the workforce planning process in professional service firms. We will present results from the model which show the impact of skill mix, skill capability levels and cross training on key performance metrics such as project completion rates, staff utilization and profit. We show how extensions to our base model can incorporate issues such as a rolling planning horizon approach and variable project start dates. 2 - Simulation -Optimization For Strategic Workforce Planning Manuel Laguna, University of Colorado Boulder, laguna@colorado.edu OptForce is a data analytics tool for workforce planning. We describe how company data is used to build a simulation model of a workforce. This model is then used for what-if analysis and optimization. The techniques associated with this application of simulation-optimization are also discussed. 3 - Talent Optimization For The Knowledge Economy Cipriano A. Santos, HP Enterprise, cipriano.santos@hpe.com Allocating the right talent for the right job at the right time, location, and cost is critical for the operational efficiency of Professional Services Organizations. In this talk we present a hierarchical planning approach for labor resources allocations

4 - Diameter-constrained Lambda-edge-connected K-subgraph Problem Yongying Zhou, University of Arizona, yongyingz@email.arizona.edu

In this talk, we study the diameter-constrained lambda-edge-connected k-subgraph problem, or the DC (k, lambda)-subgraph problem. Besides the requirements on the number of the vertices and edge connectivity, the subgraph has a diameter limit. This problem is a generalization of (k, lambda)-subgraph problem and diameter constrained minimum spanning tree. Commodity flow- based and hop constrained formulations are established, which are both integer programming (IP) formulation. Additionally, numerical experiments are performed to compare all proposed IP formulation. MD12 104B-MCC Mixed Integer Programming Formulations and Applications Sponsored: Optimization, Integer and Discrete Optimization Sponsored Session Chair: Juan Pablo Vielma, Massachusetts Institute of Technology, Cambridge, MA, United States, jvielma@mit.edu 1 - Small Independent Branching Formulations For Unions Of V-polyhedra Joey Huchette, Massachusetts Institute of Technology, Cambridge, MA, United States, huchette@mit.edu, Juan Pablo Vielma We present a framework for constructing small, strong mixed-integer formulations for disjunctive constraints. Our approach is a generalization of the logarithmically-sized formulations of Vielma and Nemhauser for SOS2 constraints, and we offer a complete characterization of its expressive power. We apply the framework to a variety of disjunctive constraints, producing novel, small, and strong formulations for outer approximations of multilinear terms, generalizations of special ordered sets, piecewise linear functions over a variety of domains, and collision avoidance constraints. 2 - Embedding Formulations For Unions Of Convex Bodies Juan Pablo Vielma, MIT, jvielma@mit.edu In this talk we extend to the non-polyhedral setting a systematic procedure to construct non-extended ideal formulations for unions of polyhedral introduced in Vielma (2015). Using geometric tools from the study of Minkoswki sums we show that the procedure can be used to recover and extend several special purpose non-extended ideal formulations. We also illustrate how the tools can be used to prove that no polynomially constrained non-extended ideal formulation can be obtained for some simple convex quadratic sets. 3 - Winning Daily Fantasy Sports Hockey Contests Using Integer Programming Scott Hunter, MIT, dshunter@mit.edu, Juan Pablo Vielma, Tauhid Zaman We present an integer programming (IP) approach to winning daily fantasy sports hockey contests which have top heavy payoff structures. Our approach incorporates publicly available predictions into a series of IPs that compute optimal lineups. We find that the produced lineups perform well in practice and are able to come in first place in contests with thousands of entries. We also show through simulations how the profit margin varies with various parameters. Our approach can easily be extended to other sports, such as American football and baseball. 4 - Smart Grids Observability Using Bilevel Programming Claudia D’Ambrosio, Ecole Polytechnique, LIX CNRS (UMR7161), Palaiseau, France, dambrosio@lix.polytechnique.fr Monitoring an electrical network is an important and challenging task. Phasor measurement units (PMU) are devices that can be used for a state estimation of this network. We consider a PMU placement problem and propose two new approaches to model this problem, which take into account a propagation rule based on Ohm’s and Kirchoff’s laws. First, we describe the natural binary linear programming model based on an iterative observability process. Then, we remove the iteration by reformulating its fixed point conditions to a bilevel program, which we solve with a tailored cutting plane algorithm. Finally, we show computational evidence of the effectiveness of our method.

MD14 104D-MCC IAAA and Sygenta Reprise

Sponsored: Analytics Sponsored Session Chair: Tarun Mohan Lal, Mayo Clinic, 1, Rochester, MN, 12345, United States, mohanlal.tarun@mayo.edu

MD15 104E-MCC Network Modeling and Inference Sponsored: Artificial Intelligence Sponsored Session

Chair: Adel Javanmard, Assistant Professor, University of Southern California, Bridge Hall, 3670 Trousdale Parkway, Los Angeles, CA, 90089, United States, ajavanma@marshall.usc.edu 1 - Co-clustering Of Non-smooth Graphons David Sungjun Choi, Carnegie Mellon University, davidch@andrew.cmu.edu Theory is becoming known for community detection and network clustering; however, the results assume an idealized model that is unlikely to hold in many settings. Here we consider exploratory co-clustering of a bipartite network whose nodes are sampled from an arbitrary population. This is equivalent to assuming a nonparametric generative model known as a graphon. We show that clusters found in the data by any method will extend to the population, or equivalently that the estimated blockmodel approximates a blocked version of the generative graphon, with error bounded by n^{-1/2}. Analogous results are also shown for degree-corrected and random dot product graph models.

212