Abstract: For complex functions with high dimensions，a new chaotic mutation differential evolution algorithm（CNDE） combined with niche is presented.Niching methods are introduced to improve the effective and independent searching.And chaotic mutation，which behaves well in local searching，is improved in this paper.In order to balance global and local search ability，convergence speed is fastened，and premature is avoided.The cross rate is automatically updated according to the generation.The experimental results show that the new algorithm not only has great advantage of convergence，but also can avoid the premature convergence problem effectively.

摘要： 针对高维复杂函数的优化问题，提出了基于小生境的混沌变异差分进化算法（CNDE）。算法结合小生境策略，使子种群高效独立地进行搜索，并引入混沌变异进行精细的遍历搜索，在运行中根据迭代次数自动地调整交叉概率因子从而使搜索的初始阶段提高种群多样性，而在搜索后期加强局部搜索能力。对3种经典函数的测试表明，新算法不仅具有很强的全局搜索能力，而且能有效避免早熟收敛问题。