Abstract: Modified differential evolution algorithm can not make effective use of knowledge about evolutionary information, and traditional cultural algorithm converge slowly because only mutation operation is adopted in population space. To solve these problems, a new hybrid optimization algorithm was proposed based on cultural algorithm and modified differential evolution algorithm. It was applied to constraint solving. Simulation tests were performed based on benchmark functions and the production scheduling problem of butene alkylation. The results indicate that the proposed algorithm is practicable and effective. Compared with other algorithms, it is superior in optimizing efficiency and results, and reduces the computational cost.

摘要： 改进差分进化算法不能有效利用进化过程中的知识，传统文化算法进化后期收敛速度较慢。针对这些问题提出一种基于文化算法和改进差分进化算法的混合算法，并将这一算法应用于约束求解问题。对基准函数和丁烯烷化生产调度问题进行仿真，结果表明该混合算法具有较好的实用性和稳健性，在寻优效率和优化结果方面都优于与之比较的算法，并降低了计算量。

关键词: 文化算法, 差分进化算法, 信念空间, 约束优化, cultural algorithm, differential evolution, belief space, constrained optimization