为寻求复杂多峰函数的全局最优解问题，提出了新型混合算法。该算法由带共享函数的遗传算法、移民技术、聚类算法和改进的Ｐｏｗｅｌ算法组成。由于上述算法的有机配合，提高了混合算法的全局和局部搜索能力。油藏系统应用实例和仿真实例证明了算法的有效性 In order to solve the global optimal solution of complex multimodal functions, a new hybrid algorithm is proposed. The algorithm consists of genetic algorithm with shared function, immigration technique, clustering algorithm and improved Powel algorithm. Due to the organic cooperation of the above algorithms, the global and local search capabilities of the hybrid algorithm are improved. The reservoir system application examples and simulation examples demonstrate the effectiveness of the algorithm