Using Improved Genetic Algorithm to Solve Game Problem
Game machine is an important research field of artificial intelligence. An improved genetic algorithm is put forward in the paper, and it can solve the problem of the machine game effectively. Phase-out factor will be introduced in genetic algorithms. By
- PDF / 74,440 Bytes
- 7 Pages / 439.37 x 666.142 pts Page_size
- 6 Downloads / 225 Views
Using Improved Genetic Algorithm to Solve Game Problem Qiujie Zhang
Abstract Game machine is an important research field of artificial intelligence. An improved genetic algorithm is put forward in the paper, and it can solve the problem of the machine game effectively. Phase-out factor will be introduced in genetic algorithms. By analyzing the value of individual fitness function and eliminating the individual whose value is low, we can increase the proportion of good genes and speed up the convergence rate of groups. Data simulation results have show that the improved algorithm has a strong ability in the area of local searching and global optimization. It can improve the accuracy and speed of the solution. Keywords Genetic algorithm • Game • Phase-out factor
55.1
Introduction
We can use the search tree method to solve game machine problem, the search tree method is a kind of game tree traversal, with the search depth increases, the computer of calculate the volume will is increasing exponentially, this is all the computers cannot afford, and even further increase the speed of the computer will not help. Thus, the search tree method to resolve the game machine problem usually only a very limited search to depth, according to the limited depth of the situation to determine the response of the pros and cons, often without a good deal results. In the genetic algorithm [1], the biological evolution is look as a long-term optimization process. Use the idea of biological evolution to solve some of the more complex issues. Biological population based search mechanism, emphasizing the exchange of information between individuals, with strong versatility and strong
Q. Zhang (*) College of Science, Heilongjiang Institute of Science and Technology, Harbin, Heilongjiang, China e-mail: [email protected] S. Zhong (ed.), Proceedings of the 2012 International Conference on Cybernetics 427 and Informatics, Lecture Notes in Electrical Engineering 163, DOI 10.1007/978-1-4614-3872-4_55, # Springer Science+Business Media New York 2014
428
Q. Zhang
ability of global optimization. Improved genetic algorithm [2] to improve search speed, we add eliminated Factor, so the fitness function value of the individual is small will be eliminated, thereby improving the convergence rate of Biological population, in order to achieve the purpose of improving accuracy and speed.
55.2
Improved Genetic Algorithm Idea
Genetic algorithm is a simulation of biological genetic approach [3]. In the genetic algorithm, the first randomly select some initial solution, the aim of genetic algorithms is the optional operation of the individuals according to their adaptability of the environment. Completion of each selection must be based on a certain probability of crossover and mutation operating, the new populations be selection— crossover—mutation loop operation. According to the theory of natural selection, genetic algorithms can solve the problem, close to the optimal solution to achieve the desired accuracy. Based on genetic algorithm, each time befor
Data Loading...
