Particle Swarm Optimization and priority representation
MetadataShow full item record
In this thesis I examine the poor performance of Discrete Particle Swarm Optimization when applied to forest planning, an optimization problem in which the goal is to maintain an even flow of timber from a forested area. I consider an alternative priority representation that encodes a permutation or ordering of plan elements in real numbers to improve the handling of constraints. I also examine its applications to two other constrained optimization problems, n-queens and snake-in-a-box, in order to show how it handles different kinds of problems. I find that priority representation is a useful tool for optimization within constraints.