Informs Annual Meeting 2017

TB36

INFORMS Houston – 2017

TB38

6 - Effect of CSR on Product Differentiation in the Presence of Cost Advantage Samar K. Mukhopadhyay, Professor, University of Wisconsin- Milwaukee, 3200 N.Maryland Ave., Milwaukee, WI, 53211, United States, samar@uwm.edu, Guangliang Ye, Lu Xu This paper links a firm’s CSR strategy with its product differentiation strategy when it competes within another firm in a Hotelling type product differentiation line. The CSR firm maximizes a convex combination of its profit and a form of utility function, while the non-CSR firm maximizes its profit only. The CSR firm is also assumed to have a technological advantage that reduces its production cost. The interaction of the effects of both the extent of CSR is considered along with a scenario of asymmetric information. Our main results show that the degree of product differentiation is reduced when CSR is practiced. On the other hand, product differentiation increases with production cost advantage.

352C MCDM Challenges in Highly-Constrained Problems Sponsored: Multiple Criteria Decision Making Sponsored Session Chair: Stephen Henry, Sandia National Labs, Albuquerque, NM, 87109, United States, smhenry@sandia.gov 1 - Increasing Stakeholder Confidence in Multi-criteria Decision Analysis Lucas Waddell, Sandia National Laboratories, 6401 Santa Monica Avenue NE, Apt 2085, Albuquerque, NM, 87109-4166, United States, lucas.waddell44@gmail.com Optimization techniques are increasingly being used to inform high-consequence, national-scale decision making. These problems are typically multi-objective, nonlinear, and highly constrained, necessitating heuristic solution methods. Lack of optimality conditions, simplifying modeling assumptions, and imperfect data often leave stakeholders asking the gut-wrenching question, “How can I be confident that the analysis is leading to well-informed decisions?” This talk will explore practical ways that we have increased analyst and customer confidence in our modeling and optimization techniques. 2 - Addressing Inconsistent Trade-Space Exploration in Constrained Genetic Algorithms Marissa Ballantine, Sandia National Labs, Albuquerque, NM, United States, mdballa@sandia.gov The Whole System Trades Analysis Tool (WSTAT) employs a multi-objective genetic algorithm to identify Pareto optimal system designs that also adhere to the physical and operational constraints of the system. Systems with complex constraints can yield solutions that fall into one of several “system architectures” (defined by the manner the constraints are satisfied). Unfortunately, genetic operators can have difficulty jumping from one architecture to another, resulting in solution sets that are inconsistent from run to run. The talk presents research for automatically identifying multiple system architectures and ensuring the genetic algorithm represents each one fairly and consistently. 3 - Integrating System Design and Portfolio Optimization: A Computational Study Matthew Hoffman, Sandia National Laboratories, 1624 Singletary Dr. NE, Albuquerque, NM, 87112, United States, mjhoffm@sandia.gov, Stephen Henry, Lucas Waddell, Frank Muldoon Optimal system design should consider the portfolio into which the system will be integrated, but representing portfolio constraints in a multi-objective metaheuristic is typically intractable. We will review previous work that described a method for incorporating multiobjective system design tradeoff information into a mixed-integer portfolio optimization, using a convex hull representation of the system design Pareto set. We will discuss refinements to the method and provide computational results comparing this method versus a typical approach of including all system design Pareto solutions in the portfolio optimization and limiting to a choice of one design via SOS1 constraints. 4- Multi-objective Optimization for Task Mode Selection and Scheduling Utilizing a Genetic Algorithm Alexander Dessanti, Sandia National Laboratories, P.O. Box 5800, MS.1188, Albuquerque, NM, 87185-1188, United States, adessan@sandia.gov Schedule Management Optimization (SMO) is a new analytic capability that utilizes a multi-objective genetic algorithm to identify optimal combinations of activities to complete a set of tasks for a project as well as the optimal schedule for each set of selected activities. SMO is flexible and can account for various constraints such as precedence relationships between activities, resource availability, and fixed dates for certain activities. Focus will be on introducing the SMO capability, pros and cons of using a genetic algorithm for this type of problem, and how constraints are handled.

TB36

351F Diversification Invited: INFORMS Special Sessions Invited Session

Chair: Michael P Johnson, University of Massachusetts-Boston, Department of Public Policy & Public Aff, 100 Morrissey Boulevard, Boston, MA, 02125-3393, United States, michael.johnson@umb.edu Co-Chair: Ruben Proano, Rochester Institute of Technology, Gleason College of Engineering, 81 Lomb Memorial Drive, Rochester, NY, 14623, United States, rpmeie@rit.edu 1 - Strengthening the Profession Through Diversity, Inclusion and Equity: Best Practices and Distinguished Practitioners Michael P.Johnson, University of Massachusetts-Boston, Department of Public Policy & Public Aff, 100 Morrissey Boulevard, Boston, MA, 02125-3393, United States, michael.johnson@umb.edu This session will introduce INFORMS conference participants to the mission and goals of the diversity, inclusion and equity committee. Panelists will describe the state of diversity and inclusion at INFORMS and other professional societies; at organizations whose employees and members participate in INFORMS and other professional societies; and the community of practice associated with operations research, management science and analytics, other STEM fields and related disciplines. Speakers and audience members will explore values, strategies and tactics that may enable INFORMS members of diverse backgrounds and experiences to achieve professional success, and that may enable organizations to produce greater social impact through best practices in diversity and inclusion. 2 - Panelist Ruben Proano, Rochester Institute of Technology, Gleason College of Engineering, 81 Lomb Memorial Drive, Rochester, NY, 14623, United States, rpmeie@rit.edu 3 - Panelist Deborah H. Urbanski, Director, Office of Equal Opportunity & Diversity, Nasa Johnson Space Center, Houston, TX, United States, deborah.h.urbanksi@nasa.gov 4 - Panelist Jeffrey Lowe, Co-Chair, Planners of Color Interest Group of the, Texas Southern University, Houston, TX, United States, lowejs@tsu.edu 352B Service Science Analytics Competition Award Sponsored: Service Science Sponsored Session Chair: Robin Qiu, Pennsylvania State University, 30 E. Swedesford Road, Malvern, PA, 19355, United States, robinqiu@psu.edu TB37

316