连续编码的基因表达式编程算法
发布时间：2019-06-10 14:39
【摘要】：提出一种基因连续编码的改进GEP算法。通过改造K表达式的编码规则消灭基因内区,使用更短的编码规则,尽可能利用到每位编码,保持基因多样性,改善其跳出局部最优的能力。由于充分利用基因连续编码的天然基因片断特性,不须定义多基因结构,降低算法设置的遗传算子数量,减少人工参数设置对算法的干扰。在序列推理和函数发现数据集上的实验结果表明,该算法具有更快的运算速度、较高的精度和较强的寻优能力。
[Abstract]:An improved GEP algorithm for gene continuous coding is proposed. By modifying the coding rules of the K expression, the inner region of the gene is eliminated, and a shorter coding rule is used, so that the gene diversity can be kept as much as possible, and the ability of jumping out of the local optimal can be improved. Due to the full utilization of the natural gene fragment characteristics of the gene continuous coding, the multi-gene structure is not required, the number of the genetic operators set by the algorithm is reduced, and the interference of the manual parameter setting on the algorithm is reduced. The experimental results of the sequence reasoning and the function discovery data set show that the algorithm has faster operation speed, higher precision and better optimization ability.
【作者单位】： 华南农业大学数学与信息学院;华南家禽疫病防控与产品安全协同创新中心;
【基金】：国家自然科学基金项目(71472068) 广东省科技计划基金项目(2012A020602102) 广州市科技计划基金项目(2014Y4300006)
【分类号】：Q811.4;TP18
，
本文编号：2496525
[Abstract]:An improved GEP algorithm for gene continuous coding is proposed. By modifying the coding rules of the K expression, the inner region of the gene is eliminated, and a shorter coding rule is used, so that the gene diversity can be kept as much as possible, and the ability of jumping out of the local optimal can be improved. Due to the full utilization of the natural gene fragment characteristics of the gene continuous coding, the multi-gene structure is not required, the number of the genetic operators set by the algorithm is reduced, and the interference of the manual parameter setting on the algorithm is reduced. The experimental results of the sequence reasoning and the function discovery data set show that the algorithm has faster operation speed, higher precision and better optimization ability.
【作者单位】： 华南农业大学数学与信息学院;华南家禽疫病防控与产品安全协同创新中心;
【基金】：国家自然科学基金项目(71472068) 广东省科技计划基金项目(2012A020602102) 广州市科技计划基金项目(2014Y4300006)
【分类号】：Q811.4;TP18
，
本文编号：2496525
本文链接：https://www.wllwen.com/kejilunwen/jiyingongcheng/2496525.html
最近更新
教材专著
