Multi-reserved strategy and its application in evolutionary computation

As a kind of intelligence computation method, evolutionary computation is widely applied and ceaselessly developed. Generally, it is made up of genetic algorithm, evolutionary strategies and evolutionary programming. And genetic algorithm is one of the most common ones, it has the features of easy structure and strong adaptability, achieves great success in many real fields, but it has much shortcomings such as greater computation complexity, more chance of being trapped into local states and the premature convergence. In this paper, by analyzing the deficiencies of the existing genetic operation and the essential characteristics of creature evolution, starting from the angle of improving evolution efficiency, we propose multi-reserved strategy based on intelligence evolution; Furthermore, establish a kind of genetic algorithm named by MGA; Finally, we analyze the performances of MGA by the theory of Markov chains and an example. All the results indicate that, MGA is obviously better than ordinary GA in computation efficiency and convergence performance.