Mean-adaptive Real-coding Genetic Algorithm and its Applications to Electromagnetic Optimization (Part One)

In the paper, a novel instance of the real-coding steady-state genetic algorithm, called the Mean-adaptive real-coding genetic algorithm, is put forward. In this instance, three novel implementations of evolution operators are incorporated. Those are a recombination and two mutation operators. All of the evolution operators are designed with the aim of possessing a big explorative power. Moreover, one of the mutation operators exhibits self-adaptive behavior and the other exhibits adaptive behavior, thereby allowing the algorithm to self-control its own mutability as the search advances. This algorithm also takes advantage of population-elitist selection, acting as a replacement policy, being adopted from evolution strategies.
The purpose of this paper (i.e., the first part) is to provide theoretical foundations of a robust, advanced and widely applicable instance of the real-coding genetic algorithm having a big potential of being successfully applied to electromagnetic optimization.