基于表现型的基因表达式编程解空间模型研究
发布时间：2019-02-17 19:52
【摘要】：基因表达式编程(gene expression programming,GEP)解空间模型理论对算法性能的改进有现实指导意义。公开文献对GEP解空间模型的研究较少,鲜见针对GEP表现型的理论研究。基于此,提出一种基于表现型的GEP解空间模型。首先,通过定义GEP染色体表现型高度,给出单基因染色体和多基因染色体表现型高度确定上界的定理及证明,利用GEP算法自身函数发现的能力,探索出操作符集最小目数为1或2的GEP染色体表现型高度上界计算的通项公式,以保证GEP表现型解空间模型的确定有界性与可计算性。其次,以GEP表现型高度的确定上界定理为基础,构建基于表现型的GEP解空间模型,总结GEP表现型解空间模型的性质和定理。通过进一步定义GEP表现型的完全解空间概念,对最优解在GEP表现型解空间和完全解空间中的分布特征进行探索研究,获知在完全解空间中最优解随子空间序号的增长呈大比例增加的分布特征。基于表现型空间模型知识,提出限制GEP种群搜索空间的基本思想与控制策略,利用模型知识合理地解释公开文献中多种GEP改进算法的有效性。
[Abstract]:The theory of solving space model of gene expression programming (gene expression programming,GEP) has practical significance to improve the performance of the algorithm. There are few studies on GEP solution space model in the open literature, and few theoretical studies on GEP phenotype. Based on this, a representation-based GEP solution space model is proposed. First of all, by defining the height of GEP chromosome phenotype, the theorem and proof of determining the upper bound of the height of single gene chromosome and polygene chromosome phenotype are given, and the ability of GEP algorithm to discover its own function is given. In order to ensure the boundedness and computability of the GEP phenotype solution space model, a general term formula for calculating the upper bound of the height of the GEP chromosome phenotype with the minimum number of mesh 1 or 2 of the operator set is explored. Secondly, based on the upper bound theorem of GEP representation height, the GEP solution space model based on representation type is constructed, and the properties and theorems of GEP representation solution space model are summarized. By further defining the concept of complete solution space of GEP representation type, the distribution characteristics of optimal solution in GEP representation solution space and complete solution space are studied. It is found that the growth of the ordinal number of the optimal solution subspace in the complete solution space is increased in large proportion. Based on the knowledge of representation space model, this paper puts forward the basic idea and control strategy of restricting the search space of GEP population, and reasonably explains the effectiveness of various improved GEP algorithms in open literature by using model knowledge.
【作者单位】： 黔南民族师范学院计算机与信息学院;广州大学计算机科学与教育软件学院;武汉大学软件工程国家重点实验室;
【基金】：国家自然科学基金资助项目(61170199) 贵州省科技厅联合基金项目资助(20157727;2013GZ12215)
【分类号】：Q811.4
本文编号：2425537
[Abstract]:The theory of solving space model of gene expression programming (gene expression programming,GEP) has practical significance to improve the performance of the algorithm. There are few studies on GEP solution space model in the open literature, and few theoretical studies on GEP phenotype. Based on this, a representation-based GEP solution space model is proposed. First of all, by defining the height of GEP chromosome phenotype, the theorem and proof of determining the upper bound of the height of single gene chromosome and polygene chromosome phenotype are given, and the ability of GEP algorithm to discover its own function is given. In order to ensure the boundedness and computability of the GEP phenotype solution space model, a general term formula for calculating the upper bound of the height of the GEP chromosome phenotype with the minimum number of mesh 1 or 2 of the operator set is explored. Secondly, based on the upper bound theorem of GEP representation height, the GEP solution space model based on representation type is constructed, and the properties and theorems of GEP representation solution space model are summarized. By further defining the concept of complete solution space of GEP representation type, the distribution characteristics of optimal solution in GEP representation solution space and complete solution space are studied. It is found that the growth of the ordinal number of the optimal solution subspace in the complete solution space is increased in large proportion. Based on the knowledge of representation space model, this paper puts forward the basic idea and control strategy of restricting the search space of GEP population, and reasonably explains the effectiveness of various improved GEP algorithms in open literature by using model knowledge.
【作者单位】： 黔南民族师范学院计算机与信息学院;广州大学计算机科学与教育软件学院;武汉大学软件工程国家重点实验室;
【基金】：国家自然科学基金资助项目(61170199) 贵州省科技厅联合基金项目资助(20157727;2013GZ12215)
【分类号】：Q811.4
【相似文献】
相关期刊论文 前1条
1 郝冬梅,阮晓钢;空间模型对单次运动相关脑电的分析[J];中国生物医学工程学报;2005年01期
，本文编号：2425537
本文链接：https://www.wllwen.com/kejilunwen/jiyingongcheng/2425537.html
最近更新
教材专著
