混合策略改进的草原土拨鼠算法及其应用

Hybrid strategy improved prairie dog algorithm and its application

针对草原土拨鼠算法收敛速度慢、易陷入局部最优、搜索与开发阶段过渡生硬等问题,提出一种混合策略改进的草原土拨鼠算法(hybrid strategy improved prairie dog optimization, HIPDO)。Piecewise混沌映射被引入以生成均匀性好的初始草原土拨鼠种群;自适应概率阈值过渡策略调整草原土拨鼠算法不同阶段的划定条件;天敌躲避策略增强草原土拨鼠躲避捕食者的能力;次优个体引导策略降低了算法被带入局部极值的风险。利用16个基准测试函数进行实验,结果表明HIPDO比对比的9个算法在更多类型的函数里面有优越性;最后将HIPDO应用到齿轮设计问题和比例-积分-微分(proportion integration differentiation, PID)参数优化问题当中,结果进一步证明了HIPDO是一种求解精度高、适用范围广的实用算法。

Aiming at the problems of slow convergence speed,easy to fall into local optimum,and excessive rigidity in search and development stage of prairie dog optimization algorithm,a hybrid strategy improved prairie dog optimization algorithm(HIPDO)is proposed.Piecewise chaos mapping is introduced generate a homogeneous initial population of prairie dogs.The adaptive probability threshold transition strategy adjusts the delimitation conditions for different stages of PDO.Natural enemy avoidance strategy enhances prairie dogs’ability to avoid predators.Suboptimal individual bootstrap strategy reduces the risk of the algorithm being carried to local extremes.16 benchmark functions are used for the experiment,the results show the superiority of the HIPDO over the 9 comparison algorithms for a wider range of types of functions.Finally,HIPDO is applied to the gear design problem and proportion integration differentiation(PID)parameter optimization problem,the experimental results further verify that HIPDO is a practical algorithm with high solution accuracy and wide range of application.