关键词: 多目标优化, 烟花算法, 差分算法, 混合优化算法

Abstract: Fireworks algorithm is an effective heuristic swarm intelligence algorithm proposed recently. But the basic fireworks algorithm can only solve the single objective problem, and it lacks information communication between individuals. The useful information is underutilized in the process of evolution. In order to improve the comprehensive performance of the fireworks algorithm and make it be used in solving multi-objective optimization problems (MOPs), this paper puts forward a kind of multi-objective hybrid optimization algorithm of fireworks and differential guided by evolution information (OHMFWDE). The Pareto frontier individual evolutionary information is used to guide the evolution of the population to accelerate the algorithm convergence speed. Crossover operator and mutation operator of differential algorithm are used instead of the original Gaussian mutation operator in fireworks algorithm to strengthen the exchange of information between individuals. Experimental results show that MOHFWDE is an effective multi-objective optimization algorithm which has good convergence, diversity and approximation compared with other algorithms.

Key words: multi-objective optimization, fireworks algorithm, differential algorithm, hybrid optimization algorithm