关键词: 萤火虫算法, 邻域结构, 冯诺依曼结构, 自适应交叉策略

Abstract: Aiming at the problems of Firefly Algorithm（FA）, such as high complexity, difficulty in optimizing high-dimensional function and easily to fall into local minimum, a new algorithm based on topology improvement and crossover strategy is proposed. The algorithm uses von Neumann topology structure to simulate the neighborhood structure between fireflies, which strengthens the global search ability and reduces the computational complexity. At the same time, adaptive crossover strategy is introduced in this algorithm. The crossover probability is adjusted dynamically according to the diversity of fireflies, which enhances the ability of fireflies to jump out of the local optimum. The simulation results of eight standard test functions indicate that the improved firefly algorithm has higher convergence accuracy and stability than the standard firefly algorithm.

Key words: Firefly Algorithm（FA）, neighborhood structure, von Neumann structure, adaptive crossover strategy