Abstract: This paper proposes a novel genetic algorithm for numerical optimization problems with continuous variables, in which a hybrid crossover operator is designed to improve the fitness of individuals by means of combining traditional crossover operators with a new optimization technique, as well as a modified fitness function. Moreover, it is demonstrated that the new algorithm is globally convergent. The numerical results show the proposed algorithm more efficiently enhance GA than do the methods in the literatures on these test functions.

Key words: Genetic algorithms, Numerical optimization problems, Hybrid crossover operator, Fitness functions

摘要： 本文将传统遗传算法中的杂交算子与一种新设计的优化方法相结合，提出了一种能改善种群中个体适应度的混合杂交算子，并通过修正适应度函数给出了一种新的求解连续型数值优化问题的遗传算法，并证明了其全局收敛性。数据试验表明，该算法对这些测试函数的结果优于文献中的方法

关键词: 遗传算法, 数值优化问题, 混合杂交算子, 适应度函数