Publication Date
2006
Document Type
Thesis
Committee Members
Raymond Hill (Advisor)
Degree Name
Master of Science in Engineering (MSEgr)
Abstract
Real-world complex optimization problems are difficult to solve. Agent-based optimization approaches have proved useful in solving a wide variety of problems including optimization problems. Agent-based techniques can be used in military planning for solving allocation problems such as the weapons to targets assignment problem. Classical methods like linear programming (LP) have been used for solving weapons to targets assignment problems. LP approaches provide optimal solutions quickly, but in real-time planning when there are minor changes to input, LP exhibits widely varied solutions. This can be a problem in practice. This research study considers two agent-based optimization approaches, the Stable Marriage Algorithm (SMA) and the Ant-Colony Optimization (ACO) algorithm, for solving the weapons to targets assignment problem. In real-time defense planning and re-planning scenario, the effect of the input data changes on the solutions provided by SMA and ACO is observed. An interactive tool is developed in Visual Basic 6.0 for performing the assignment of weapons to targets using either of the agent-based optimization algorithms. An empirical analysis for determining the best parameter settings for finding good solutions for ACO algorithm is carried out. The performance of SMA and ACO is compared in terms of solution quality and persistence characteristics. Results indicate better performance of SMA than ACO in terms of persistence. In terms of solution quality, ACO provides solutions with lower assignment cost values than SMA.
Page Count
58
Department or Program
Department of Biomedical, Industrial & Human Factors Engineering
Year Degree Awarded
2006
Copyright
Copyright 2006, all rights reserved. This open access ETD is published by Wright State University and OhioLINK.
