Distinguishing Genotype and Phenotype in Genetic Programming

This paper introduces GADS, a technique for genetic programming where the genotype is distinct from the phenotype. The GADS genotype is a list of integers representing productions in a syntax. This is used to generate the phenotype, which is a program in the language defined by the syntax. Syntactically invalid phenotypes cannot be generated, though there may be phenotypes with residual nonterminals. GADS can be implemented on a traditional genetic algorithm. The paper describes an experiment to decide whether GADS is feasible and to explore the effect of some variables on its performance. The results show that GADS can be more efficient than traditional tree-based genetic programming. Contents 1 Introduction 2 Principles of GADS 3 Experimental design 4 Experimental results 5 Comparisons 6 Future research 7 Conclusions Bibliography 1 Introduction This section explains the background in genetic algorithms against which the GADS development has taken place. Genetic programming (GP) was i..