Quantum-Behaved Particle Swarm Optimization Using Q-Learning

Abstract

Quantum-behaved particle swarm optimization (QPSO) has shown excellent performance in solving optimization problems which inspired by analysis from particle swarm optimization (PSO) and quantum mechanics. In QPSO, the only parameter contraction-expansion coefficient β is vital to the performance of algorithm. This paper employs Q-Learning strategy and presents a novel parameter control method to improve QPSO performance. Then the empirical studies on a suite of well-known benchmark functions are to be performed to test performance. Finally, a further performance comparison between the proposed algorithm and other parameter control methods of QPSO are listed and the simulation results show the efficiency of the proposed QPSO with novel adaptive strategies.