Informs Annual Meeting Phoenix 2018

INFORMS Phoenix – 2018

MB06

5 - QSDPNAL: A Two-phase Augmented Lagrangian Method for Convex Quadratic Semidefinite Programming Xudong Li, Princeton University, NJ, United States, Defeng Sun, Kim-Chuan Toh In machine learning, statistics, and other areas, numerous interesting problems can be modeled in the form of convex composite quadratic conic programming. In this talk, we use the convex quadratic semidefinite programming (QSDP) problem as an example to introduce a two-phase augmented Lagrangian method (ALM), called QSDPNAL, for solving QSDP with linear equality and inequality constraints. Under mild conditions, we are able to establish the rate of convergence of the proposed algorithm and prove the R-(super)linear convergence of the KKT residual. Extensive numerical results show that our algorithm ishighly efficient and robust in obtaining accurate solutions. n MB06 North Bldg 122C Optimization and System Analysis Sponsored: Optimization/Computational Optimization and Software Sponsored Session Chair: Les Servi, The MITRE Corporation, Bedford, MA, 01730-1420, United States 1 - Comparison of Solution Methods in Evaluating a System of Risk Assessments Mark Gallagher, Air Force, Washington, DC, United States, John Lepird, Daniel Fenn, Shane Hall Military operations are a synthesis of capabilities, challenging integrated risk assessment. We establish a framework to integrate capabilities using a common risk metric, proposing a novel approach to establishing relationships between them, and deriving models aggregating risks and dependencies. This framework has informed billions of dollars of Air Force investments. 2 - Offensive and Defensive Naval Weapon Assignment to Swarming Threats Connor S. McLemore, Navy, Washington, DC, United States We propose an automated decision aid capable of generating offensive and defensive engagement profiles for naval weapons. It allows the efficient pairing of multiple weapon systems to multiple swarming threats operating in multiple domains by providing the operator with recommended weapon-target pairings based on current capabilities and threat profiles. The model consists of pre- processing algorithms and reward-based mixed-integer programming models that takes as inputs the available weapon system capabilities and target information and outputs a recommended engagement profile. 3 - A System-of-Systems Methodology for Analysis of the Behavior of Complex System Dan DeLaurentis, Purdue University, West Lafayette, IN, United States, Cesare Guariniello System Operational Dependency Analysis (SODA) is a methodology to assess the effect of dependencies between constituent elements of complex systems. The analysis, based on a parametric model of the behavior of the elements, is domain- agnostic and provides identification of critical areas, study of cascading effect of disruptions, and assessment of features of the whole system, such as robustness and resilience. SODA can be used to compare different systems architecture for decisions in early design phase, and to assess vulnerabilities and potential improvements of existing architectures. 4 - Mission Dependency Analysis with Physical Network Relationship Les Servi, The MITRE Corporation, M/S M230, 202 Burlington Road, Bedford, MA, 01730-1420, United States A functional dependency network is a useful framework to evaluate alternative decisions under a constrained budget. It captures the impact of decisions on the hierarchy of goals and sub-goals required to achieve a mission. Motivated by cyber security considerations, this talk will describe the use of mixed integer linear programming to optimize the defending or attacking such networks but also taking into account the physical network associated with the functional network.

n MB07 North Bldg 123 Approaches in Security Problems Sponsored: Optimization/Network Optimization Sponsored Session Chair: N. Orkun Baycik, Shenandoah University, Winchester, VA, n/a, United States 1 - Computing the Vulnerability of Multi-hop Wireless Networks: A Search for the ôBest Interference Model Hugh Medal, Mississippi State University, P.O. Box 9542, MSU, MS, 39762, United States, Randy K. Buchanan This paper studies the problem of placing a set of jammers in 3-D space in order to minimize the throughput of a wireless communication network. The main goal of this paper is to study the effects of jamming under several different models of interference (e.g., physical, capture, protocol, interference range). Toward this end, this paper presents a mixed-integer programming (MIP) formulation and branch-and-cut procedure for the jammer location problem under several models of interference as well as a study of the effect of different interference models on runtime and solution tractability. 2 - Fishing and Trafficked Labor: An Examination of Policies to Address the Intersection of Prosperity and Exploitation Renata Alexandra Konrad, Worcester Polytechnic Institute, School of Business, 100 Institute Road, Worcester, MA, 01609, United States, Khalid Saeed, Matt Kammer-Kerwick Human exploitation in the seafood industry is a complex transnational problem jeopardizing human rights and marine ecosystems. Labor trafficking, environmental sustainability, aquaculture, and socio-economic development interact interdependently, and form a large system with multiple decision makers with conflicting goals. We present the results of a System Dynamics simulation model which incorporates resource management of fish stocks and illicit labor markets. Using this model, we examine the implications on trafficked labor of several policies including: imposing an excise tax, trafficking prevention campaigns, and increased policing. 3 - Interdicting Layered Physical and Information Flow Networks N. Orkun Baycik, Shenandoah University, Winchester, VA, United States, Thomas Sharkey, Chase Rainwater We focus on the problem of interdicting layered networks that involve a physical flow network and an information flow network. There are interdependencies between the networks which leads to a network interdiction problem with a discrete inner problem. The objective of the defender is to send the maximum amount of flow through its physical flow network whereas the attacker seeks to minimize this maximum flow. For the case where the information supply arcs are uncapacitated, we apply a multi-step, dual-based reformulation technique. We apply this technique to law enforcement efforts against illegal drug trafficking and cyber vulnerability analysis of infrastructure networks. n MB08 North Bldg 124A Connectivity and Networks Sponsored: Optimization/Network Optimization Sponsored Session Chair: Hosseinali Salemi, Oklahoma State University, Stillwater, OK, 74075, United States 1 - Connectivity and Power Domination in Graphs Logan Smith, Graduate Student, Rice University, Houston, TX, United States Power domination is related to the placements of sensors in electrical networks. Minimum power dominating sets in a graph correspond to minimal sensor arrays that are able to fully observe a given power grid. Nearby sensors can be supported with reduced infrastructure, thereby incentivizing the identification of minimum power dominating sets with induced subgraphs having a minimum number of components. Methods in power domination and connectivity constraints are adapted to solve this problem and computational experiments compare runtime performances. 2 - A Lagrangian Relaxation Upper Bound on Clique Number of Graphs Seyedmohammadhossein Hosseinian, Texas A&M University, 3131 TAMU, College Station, TX, 77843-3131, United States, Sergiy Butenko We introduce a new analytic (i.e., closed-form) upper bound on clique number of graphs. This bound is derived from a Lagrangian relaxation of an integer (linear) programming formulation of the maximum edge weight clique problem, and is characterized by degrees of vertices in the graph. We compare this bound with other analytic bounds proposed in the literature.

150