Mixed Integer Programming for the Dynamic Allocation of Detection Assets in Undersea Warfare
Abstract
Project/Abstract APPROVED FOR PUBLIC RELEASEThe research project Mixed Integer Programming for the Dynamic Allocation of DetectionAssets in Undersea Warfare will advance modeling and algorithmic capabilities of mathematicaloptimization to assist Anti-Submarine Warfare (ASW) planning.The ability to localize and neutralize adversary submarines relies upon analysis of detectionsmade and predictions of where adversary submarines are headed. Crucial to our research plan willbe leveraging an existing predictive analytic tool for ASW that exploits intelligence information,historical analysis, and tactical contact reports to generate probabilistic estimates of target locationand intended mission. The tool can project future target location even in the face of largetime gaps between contacts reports and high degrees of uncertainty over intended target mission.Our proposed research will augment this tool, utilizing its target location predictions to suggestcoordinated ASW search plans using mathematical optimization techniques.By developing, prototyping, solving, and benchmarking sophisticated optimization models,combined with algorithmic advances, we aim to assist Theater Undersea Warfare Commanders.The end result will be a usable decision-making tool to help commanders place Blue detectionassets and suggest search paths in order to achieve mission objectives. At each phase of the project,models and algorithms will be integrated into a simulation system for evaluation of the policiessuggested by the mathematical models. We also will work closely with domain-experts to ensurethe models we build are realistic enough to be readily deployable for practical operations.The success of the research project hinges on a number of key research questions that will beaddressed:1. Can scalable discrete optimization models be built that accurately account for necessaryoperational considerations while simultaneously considering many different commander objectives?2. Canthe impact of failed detection events on subsequent target probability distributions bemathematically modeled and incorporated as algenraic constraints to an optimization modelin a tractable manner?3. What level of fidelity of exogenous random factors is necessary to consider in order to buildmodels that result in robust operational search plans?4. Can practically-sized optimization models that explicitly account for both exogenous andendogenous uncertainty be solved?The overarching goal of the research project is to create mathematically-sound software toolsthat are implementable and useable as decision-making aids for Theater UnderseaWarfare officers.The project may also serve the purpose of demonstrating the utility of applying advanced mathematicaloptimization tools, such as integer programming and stochastic programming, widely innaval operations.
Document Details
- Document Type
- DoD Grant Award
- Publication Date
- Apr 14, 2023
- Source ID
- N000142312338
Entities
People
- Jeffrey Linderoth
Organizations
- Office of Naval Research
- United States Navy
- University of Wisconsin System