【摘要】 在求解全局优化问题时,通常免疫算法、进化扩散算法分别在局部搜索和全局搜索方面表现较弱。针对这一情况,基于免疫和进化扩散算法,提出了一个免疫-进化扩散算法。该算法结合了免疫和进化扩散两种算法的优点,一方面通过引入基于共享机制的小生境算法,保持了群体的多样性,另一方面通过提出一种步长参数动态调整策略,提高了算法效率。实验结果表明,在给定精度下,该算法的效率和稳定性都明显优于Tsui的进化扩散算法和Ingber的自适应模拟退火算法。最后对步长参数动态调整策略进行了分析。更多还原

【Abstract】 In solving global optimization,immune algorithm is usually weak in local search,while evolutionary diffusion optimization is weak in global search.To overcome these shortcomings,an immune-evolutionary-diffusion optimization algorithm was proposed.This algorithm combines the advantages of the immune algorithm and evolutionary diffusion optimization.On one hand,it retains the diversity of colony through importing the niche algorithm;on the other hand,it improves the efficiency of the algorithm by proposing a strategy of dynamically adjusting the parameter of step size.Experimental results show that,regarding the efficiency and stability for a given precision,this algorithm performs better than Tisui’s evolutionary diffusion optimization and Ingber’s adaptive simulated annealing algorithm.The strategy of dynamically adjusting the parameter of step size was analyzed.更多还原