混合策略改进的蝴蝶优化算法

针对蝴蝶优化算法存在的求解精度低、易陷入局部最优等缺陷,提出混合策略改进的蝴蝶优化算法.首先,利用Circle映射初始化蝴蝶个体的位置,增加初始个体的多样性;其次,在局部搜索阶段利用动态切换概率控制改进正弦余弦算法与蝴蝶优化算法的转换,充分利用少量的蝴蝶个体,增强算法的局部开发能力;然后,在全局和局部位置更新处引入自适应余切权重系数,控制蝴蝶个体下一代的移动方向和距离,提高算法的收敛速度和精度;最后,引入逐维变异策略,对全局最优位置变异,引导种群向最优位置进化,避免陷入局部最优.对八个基准函数进行仿真实验,结果表明,改进算法的收敛性能更佳,与其他改进算法相比具有一定的竞争力.

Mixed strategy to improve butterfly optimization algorithm

Aiming at the defects of butterfly optimization algorithm, such as low accuracy and easy to fall into local optimum, this paper proposed an improved butterfly optimization algorithm for mixed strategy. Firstly, in order to increase the diversity of initial individuals, this algorithm used circle map to initialize the position of butterfly individuals. Secondly, in the phase of local search, it used dynamic switching probability to control the conversion between sine and cosine algorithm and butterfly optimization algorithm, so that it fully utilized a small number of butterflies to enhance the local development ability of the algorithm. Then, it introduced adaptive cotangent weight coefficients at global and local position updates to control the movement direction and distance of the next generation of butterflies and improve the convergence speed and accuracy of the algorithm. Finally, it introduced a dimensional-by-dimension mutation strategy to mutate the global optimal position, guide the population to evolve to the optimal position and avoid falling into the local optimum. The simulation results of eight benchmark functions show that the improved algorithm has better convergence performance. Compared with other improved algorithms, it has a certain competitiveness.