改进的基因表达式编程算法在演化建模中的应用

Application of improved gene expression programming for evolutionary modeling

为了克服传统基因表达式编程易早熟收敛、种群多样性难以保持、演化效率不高、拟合度不高等缺陷,给出了基于表现型的种群多样性测度,并提出了基于排挤小生境的改进基因表达式编程算法.该算法将小生境半径内的早熟个体通过罚函数排挤出去,使其它优良个体得以更大概率进化,并使各个个体之间保持一定的距离.分别对一元函数和多元复杂函数进行演化建模实验.结果表明,改进的算法能在演化过程中能保持丰富的群体多样性,能够有效避免过早收敛,具有更高的成功率、更高的收敛速度和拟合精度.

Improved gene expression programming based on crowding niche is proposed to overcome the shortcoming of the traditional gene expression programming, which is easy to premature convergence, is difficult to maintain the diversity of the population, and has low evolution efficiency and fitting accuracy. The algorithm crowd out premature individuals within niche radius through the penalty function, so that other superior individuals evolve with high probability, and each individual keep a certain distance. Through evolution modeling experiment of unary function and complex multivariate function results show that the improved algorithm can preserve population diversity, effectively avoids premature convergence, has higher success rate, faster convergence speed and higher fitting precision.