基于PRS元模型的粒子群优化算法及其应用*

针对传统粒子群算法优化黑箱模型过程中存在巨大计算开销的问题，提出一种基于PRS元模型的改进粒子群优化算法—PPSO算法。在该算法迭代过程中，构建PRS元模型，利用其最优值点辅助粒子种群的更新，此外仅选择元模型预估集中优值集的粒子进行目标函数的计算仿真。将PPSO算法与基本粒子群算法、混沌粒子群算法进行数值测试对比，并应用于模糊控制器的优化设计，仿真结果表明该算法可减少真实估值次数,提高优化搜索能力。

A PRS metamodel assisted PSO optimization and the application

To reduce huge computational overheads when solving computationally expensive black-box optimization problems by conventional particle swarm optimization algorithm, this paper proposed an improved particle swarm optimization algorithm-PPSO algorithm based on the PRS metamodel. In the iterative process of PPSO, it constructed the PRS metamodel based on sampling data, and use the optimal point of metamodel to assist the update of particle population. In addition, the algorithm only selected the promising particles to perform actual function evaluations in order to reduce the computational cost. Then this paper tested the new proposed method by several benchmark functions, compare it with basis PSO and CPSO, and then apply it into the design of fuzzy controller. Numerical tests show that the proposed algorithm can reduce the number of expensive simulations and improve the search ability.