Genetic Algorithm (GA) is a king of random and global searching optimizing algorithm , which based on Dawin's natural evolution theory and Mendel's genetics and mutation theory. The main character of this algorithm is simple, general and excellent ROBUST peculiarity. It's fit for parallel distribution and procession. GA is applied in a very big range, such as machine learning, music composing and industrial control. In particular to highly complex non-liner problem, and include NP problem. When other methods are difficult to deal with question, the GA is very fit to be used.In this paper, I first research the basic theory of GA, and analyze the essences of many theories about GA for coding, difiniting of the population size, choose, crossing and mutation. With the lines machine layout, problem, I propose a coding mode fit the character of this problem, and mutation regulations for this coding mode. The method solve the low efficience of the algorithm in using based on standard coding and standard crossing mutation regulations, and the problem that be easy converge to partial extremum. Through improving operator , this new GA can do better in fact produce to solve the machine layout problem efficiently.
|