一种多策略搜索寻优的改进鸽群优化算法

A multi-strategy search improved pigeon-inspired optimization algorithm

针对鸽群优化算法易于早熟收敛、陷入局部最优解的不足,提出了一种改进的鸽群优化算法(MSIPIO).首先,在地图和指南针算子中引入免疫算法,并提出变异因子,通过交叉变异,同时提升了算法前期找到最佳寻优方向的概率;其次,在地标算子中提出种群衰减因子和全局影响因子,克服标准鸽群优化算法后期数目衰减过快的不足,增强算法寻优能力;最后,利用模拟退火机制对次优解进行保留,有效减缓标准鸽群优化算法陷入局部最优解的问题,提高了算法获得全局最优概率.仿真结果表明,与其他5种算法相比,MSIPIO算法在收敛精度上有明显提升,并且能够有效避免陷入局部最优解.

A multi-strategy search improved pigeon-inspired optimization(MSIPIO)algorithm is proposed for the shortcomings that pigeon-inspired optimization(PIO)algorithm is easy to converge early and fall into local optimal solutions.First,the immune algorithm is introduced in the map and compass operators,and variation factors are proposed to simultaneously improve the probability of finding the best search direction upfront by the algorithm through cross-variation.Secondly,the population decay factor and global influence factor are proposed in the landmark operator to overcome the shortage of rapid decay of the number in the late stage of the standard pigeon flock optimization algorithm and to enhance the algorithm’s ability to find the best.Finally,a simulated annealing mechanism is used to preserve the suboptimal solutions,which effectively mitigates the problem of the standard pigeon flock optimization algorithm falling into local optimal solutions,thus improving the probability of the algorithm obtaining the global optimum.The simulation results show that compared with the other five algorithms,MSIPIO algorithm has a significant improvement in convergence accuracy and can effectively avoid falling into local optimal solutions.