论文部分内容阅读
在样本规模有限的情况下,为了提高算法的鲁棒优化性能,提出一种基于时变(随迭代次数变化)Sigmoid函数的鲁棒粒子群优化算法.采用拟蒙特卡罗积分方法近似估计有效目标函数,以时变Sigmoid函数为基础,设计各代各样本规模的选取概率.迭代前期,样本规模期望值较小,加快了算法探索速度;迭代后期,样本规模期望值较大,提高了算法的开发精度.标准测试函数仿真结果显示,所提出方法具有较优的鲁棒优化性能.
In order to improve the robust optimization performance of the proposed algorithm, a robust particle swarm optimization algorithm based on time-varying Sigmoid function is proposed in the case of limited sample size.The quasi-Monte Carlo integration method is used to approximate the effective target Function, the time-varying Sigmoid function is used as the basis to design the probability of selection for each sample size. In the early stage of iteration, the expected value of sample size is small, which accelerates the speed of algorithm exploration. In the latter part of iteration, the expectation of sample size is larger, which improves the precision of algorithm development The simulation results of standard test function show that the proposed method has better robust optimization performance.