Continuous Hierarchical Fair Competition Model for Sustainable Innovation in Genetic Programming

Abstract

Lack of sustainable search capability of geneticprogramming has severely constrained its application tomore complex problems. A new evolutionary algorithmmodel named the continuous hierarchical faircompetition (CHFC) model is proposed to improve thecapability of sustainable innovation for singlepopulation genetic programming. It is devised byextracting the fundamental principles underlyingsustainable biological and societal processesoriginally proposed in the multi-population HFC model.The hierarchical elitism, breeding probabilitydistribution and individual distribution control overthe whole fitness range enable CHFC to achievesustainable evolution while enjoying flexible controlof an evolutionary search process. Experimental resultsdemonstrate its capability to do robust sustainablesearch and avoid the aging problem typical in geneticprogramming.