基于差分进化基因表达式编程的全局函数优化

Global Function Optimization Based on Gene Expression Programming with Differential Evolution

为了提高基因表达式编程(Gene Expression Programming,GEP)在函数优化时的效率,将差分进化(Differ-ential Evolution,DE)引入到GEP中,提出了基于差分进化的基因表达式编程的全局优化算法DEGEPO。主要工作包括:(1)针对全局函数优化问题,根据GEP和DE的特点设计了新的基因编码;(2)设计了新的变异和交叉算子;(3)提出了DEGEPO算法并进行了算法分析;(4)实验验证了算法的有效性。相对于传统GEP,DEGEPO,优化结果精度平均提高了2～4个数量级。

To improve the efficiency in function optimization via Gene Expression Programming（GEP）, Differential Evolution（DE） was introduced into GEP. A novel algorithm called DEGEPO was proposed. The main work of this paper included （1） the gene in GEP was redesigned to adapt global function optimization; （2） novel mutation and crossover operations were applied; （3） a parameter optimization algorithm based on GEP with DE called DEGEPO was proposed and it was also analyzed; （4）experiments demonstrated the efficiency and effectiveness of DEGEPO. Compared with basic GEP, the precision of DEGEPO increased 2-4 orders of magnitude averagely.