Particle Swarm Optimization Algorithm for Agent-Based Artificial Markets

摘要

Particle swarm optimization (PSO) is adapted to simulate dynamic economic games. The robustness and speed of the PSO algorithm is compared to a genetic algorithm (GA) in a Cournot oligopsony market. Artificial agents with the PSO learning algorithm find the optimal strategies that are predicted by theory. PSO is simpler and more robust to changes in algorithm parameters than GA. PSO also converges faster and gives more precise answers than the GA method which was used by some previous economic studies.