论文部分内容阅读
针对基本混洗蛙跳算法(Shuffled Frog Leaping Algorithm,SFLA)收敛速度慢、优化精度低且易于陷入局部最优的问题,对其进行多项改进。采用随机分组策略,平衡各子群的寻优能力,保持种群多样性;打破最差蛙只向最优蛙学习的模式,引入Minkowski距离,使最差蛙借助更多同伴信息选择进化方向,增强种群适应性;针对最优蛙进化机会少,引入精英策略和变异思想更新其位置,避免陷入局部极小,加快收敛速度,最后选取合适的目标函数,将改进前后混洗蛙跳算法(Improved Shuffled Frog Leaping Algorithm,ISFLA)用于冷轧液压伺服位置自动控制(Automatic Position Control,APC)系统的PID参数整定,并将整定结果进行西门子PLC实验验证,结果表明改进后算法的有效性和高效性。
Shuffled Frog Leaping Algorithm (SFLA) is improved in many aspects because of its slow convergence rate, low optimization accuracy and easy to fall into local optimum. The strategy of random grouping was used to balance the search ability of each subgroup and maintain the diversity of the population. By breaking the model of the worst frog learning from the optimal frog, the Minkowski distance was introduced to make the worst frog choose the evolutionary direction with more companion information and enhance Population adaptability. In view of that there are few optimal frog evolutionary opportunities, we introduce elitist strategies and mutation ideas to update their positions, avoid falling into local minima and speed up the convergence rate. Finally, we select the appropriate objective function and improve the improved frog leapfrog algorithm Frog Leaping Algorithm, ISFLA) is used to adjust the PID parameters of Cold Position Hydraulic Servo Position Automatic Control (APC) system. The results of the experiment are verified by Siemens PLC. The results show the effectiveness and efficiency of the improved algorithm.