Abstract: Aiming at the drawback for the quantum optimization algorithm, a novel cooperative quantum agent optimization algorithm is proposed. The individual in the population can be viewed as the agent using quantum bit code, and the evolutionary process can be divided into three phases. The information and individual can exchange between subpopulation, the individual can also compete with each other and adjust slightly. The evolutionary can carry through in the different niche, so it can enhance the evolutionary granularity. The trait of convergence is analyzed in view of the functional analysis. The fixed point theorem is used to prove the convergence of the algorithm, and the theorem shows that the proposed algorithm can reach the satisfactory solution set. Simulation results of benchmark function demonstrate that the algorithm performs well than other algorithms, and can get better solution.

Key words: agent, quantum evolutionary algorithm, cooperative evolutionary, subpopulation, chain like