Convergence rate analysis of allied genetic algorithm
Feng Lin, Chunyan Zhou, KC Chang · 2010
To support decision making, it is important to understand the convergence property of an optimization algorithm in order to design an effective system. Genetic algorithm has been applied to many difficult optimization problems. However, it is non-trivial to analyze its convergence property. In this paper, we first introduce an allied strategy and present a parallel genetic algorithm called allied genetic algorithm (AGA). We then extend the basic Markov chain model of the general elitist selected genetic algorithm (EGA) to AGA. Finally, we present a methodology to analyze the convergence rate of AGA. The preliminary experiment results show that AGA can prevent premature convergence and increase the optimization speed.