Abstract: This paper proposes a fast Pareto genetic algorithm for searching pareto optimal solution set. It is based on a new approach for fast evaluation of fitness of individuals and a clustering based external population update scheme for maintaining population diversity and even distribution of Pareto solutions. Experiments on a set of multi-objective knapsack optimization problems shows that FPGA can obtain high-quality, well distributed non-dominated Pareto solutions with less computational efforts compared to other state-of art algorithms, it has advantages in its convergence speed and quality over the state-of-the-art SPEA algorithm.

Key words: fast Pareto genetic algorithm(FPGA), Pareto optimality, fitness fast identify algorithm, fast update algorithm(FUA)

摘要： 在适应值快速辨识算法和基于聚类排挤的外部种群快速替换算法的基础上，提出了搜索Pareto最优解集的快速遗传算法。在该算法中，IFI算法实现个体适应值的快速辨识，FUA维持种群多样度和Pareto最优解集的均匀分布性。采用FPGA算法对多种多目标0/1背包问题进行仿真优化，FPGA算法能够以较少的计算成本搜索到高精度、分布均匀、高质量的Pareto非劣解集，收敛速度和收敛准确性均优于强度Pareto进化算法(SPEA)。

关键词: 快速遗传算法, Pareto最优性, 适应值快速辨识算法, 快速替换算法

CLC Number: 

• TP18