Abstract: Monarch Butterfly Optimization（MBO） is a novel swarm intelligent optimization algorithm. Yet there are still the defects of slow convergence and easy being trapped into local optima in the MBO. In order to overcome the shortcomings of the MBO, an Improved Monarch Butterfly Optimization（IMBO） is proposed in this paper. The IMBO uses the strategy of dynamic and random dividing the population into two sub-populations at every time-step, and the butterflies in different sub-populations usually use different searching methods in order to keep the diversity of population search. Experiments are done on a set of 10 benchmark functions, and the results show that the proposed algorithm has marked advantage of global convergence property, can improve the convergence efficiency in function optimization, and is more stable when being compared with MBO and PSO algorithms.

Key words: Monarch Butterfly Optimization（MBO）, optimization, intelligent computation

摘要： 大红斑蝶优化算法（MBO）是最近提出的一种新的群智能优化算法。然而，该算法仍存在收敛速度较慢、易陷入局部最优的缺点。为克服MBO算法之不足，提出了一种改进的大红斑蝶优化算法（IMBO）。该算法采用将群体动态随机分割成两个子群体的策略，不同子群体中的大红斑蝶采用不同的搜索方法，以保持种群搜索的多样性。通过10个基准函数的仿真实验并与MBO算法以及标准PSO算法相比较，结果表明IMBO算法的全局搜索能力有了明显的提高，在函数优化中具有更好的收敛速度及稳定性。

关键词: 大红斑蝶优化算法, 优化, 智能计算