Abstract: In order to improve the accuracy, speed and efficiency of the Hierarchical Cellular Genetic Algorithm (HCGA) in solving complex problem, in this paper, a new polycentric HCGA on the basis of HCGA and with reference to the central city theory of western economics was proposed. The new algorithm chose a few individuals with high fitness in population as the central cities. Individuals around the central city moved towards the center，and the optimal solution was generated from these central cities. The algorithm greatly improves populations diversity and thus the searching efficiency. The numerical simulations show that the improved algorithm is more effective for realizing the global optimization and can avoid premature effectively.

Key words: cellular genetic algorithms, central cities, high-dimension complex functions

摘要： 为提高分层元胞遗传算法在解决复杂函数优化问题时的求解精度、收敛速度和求解效率。在分层元胞遗传算法的基础上借鉴西方经济理论中中心城市思想提出了一种基于多中心城市策略的分层元胞遗传算法。该算法在进化初期选择适应度值高的多个个体作为种群进化过程中的中心城市，中心城市周围元胞空间的个体按照一定的迁移规则往中心城市迁移，全局最优解从几个中心城市中产生，这样使算法在快速收敛的同时提高了种群的多样性，从而避免落入局部最优。对几个高维的复杂函数优化问题进行了仿真验证，实验结果表明改进的算法无论在收敛速度上还是解的精度上都有较好的效果。

关键词: 元胞遗传算法, 中心城市, 高维复杂函数