基于内嵌基因表达式编程的函数优化

Function Optimization Based on Embedded Gene Expression Programming

基因表达式编程(GEP)算法在解码时常存在未表达的基因内区,在解决函数优化问题时存在缺陷,使得对简单函数的优化性能不如遗传算法(GA),而对复杂函数优化收敛速度较慢。为了改善基因表达效率和提高优化性能,做了下列工作:提出了新的基因解码方法,形成了内嵌基因表达式编程算法EGEP;设计了适合优化问题的个体编码方案;分析了个体的表达空间。实验表明,EGEP对简单函数优化的性能优于传统遗传算法;EGEP提高了对复杂函数的优化能力,即使在运行辈数降低200倍时,得到的性能仍然优于传统GEP和遗传算法。

The Gene Expression Programming（ GEP） usually exists some un-expressed introns,the performance may be lower than GA in simple function optimization and the speed is un-satisfied to complicated optimization task . To improve the expression efficiency of gene space and the performance for function optimization,an evolutionary algorithm EGEP （ Embedded Gene Expression Programming） was proposed based on a new decoding method. A new coding method for individual was designed which was suited for function optimization. And the expression space of individual was analyzed. Experiments showed that EGEP is superior to GA in simple function optimization. Even if the run generation reduced by 200 times,the performance of EGEP still surpasses GEP and GA in complex function optimization.