Sustainable Genetic Algorithms Basis of Relative Adaptation Strategy and Self-adaptive Learning Operator

This paper divides the genotype into two parts: genetic information and culture information. The self-adaptive learning algorithm (SALA) to sustain and enhance the diversity of population is put forward. This paper also proposes a relative adaptation strategy (RAS) which provides opportunity for individual schemata which is potential and a redundant reproduction operator (RRO) that naturally improves the quality of descendant and increases the convergence rate, then a kind of sustainable genetic algorithms based on comparative adaptation strategy and self-adaptive learning operator is formed (SGA). Testing on 11 problems with different parameters demonstrates that SGA is superior to standard genetic algorithms on the convergence rate as well as the capable of maintaining the diversity.