基于位串编码的遗传归纳逻辑程序设计  

Genetic Inductive Logic Programming Based on Bit-string Encoding

作　　者：杨新武[1] 刘椿年[1] 

机构地区：[1]北京工业大学计算机学院多媒体与智能软件技术北京市重点实验室,北京100022

出　　处：《计算机工程》2009年第2期13-14,20,共3页Computer Engineering

基　　金：国家自然科学基金资助重大项目(60496322);北京市组织部优秀人才基金资助项目(20051D0501508)

摘　　要：归纳逻辑程序设计是基于一阶逻辑的数据挖掘新方法。一阶规则挖掘是目标谓词和背景知识谓词对应的各种原子的复杂组合优化问题。该文根据Occam’s razor原理提出原子的位串编码,设计相应的遗传算子,基于sequential covering策略提出采用遗传算法作为搜索策略的遗传归纳逻辑程序设计算法GILP。在连通图问题和gcd问题上验证算法的可行性。Inductive logic programming is a new method for data mining, which is based on the first-order logic. The first-order rule mining is the complicated combination optimal problem of atoms constructed by target predicate and background predicates. This paper proposes the bit-string encoding of first-order rule based on the Occam＇s razor principle, designs the corresponding genetic operators, and presents the algorithm of Genetic Inductive Logic Programming（GILP） based on the sequential covering strategy and adopts genetic algorithm as the search method. It validates the feasibility of GILP algorithm through the connectivity of graphic and gcd problem.

关 键 词：遗传算法 归纳逻辑程序设计 位串编码 

分 类 号：TP181[自动化与计算机技术—控制理论与控制工程]