Artificial Life Algorithm With Predater And Application In System Recognition

Artificial Life Algorithm (ALA) is belonged to evolution algorithm with group intelligent. The system initializes one group of random result, and reaches to the optimum result after recycling. Comparing to other artificial intelligence algorithms such as GA, ALA has a simple conception, is easy to design and the parameter system is robust. Now Artificial Life Algorithm has aroused much attention, and may become another research focus. Artificial Life Algorithm is just in a start time, and there are many problems to be solved.In this thesis, Standard Artificial Life Algorithm program is realized. For improving convergent speed and precision, predator strategy, which is come from biological point of view, is introduced and the new algorithm has been designed. The new algorithm's ability is proved by optimizing the elementary test function. Artificial select strategy is abstracted in this thesis and the Two Level Algorithm is designed to reserve the relatively optimization areas. According to the proof-test, the Two Level Algorithm is better than standard algorithm in the ability on exploitation and exploration, which gives us a better way to study such problem and enlarges the application domain. The parameter system is analyzed and the basic rule is discussed in this thesis. Facing to multidimensional problem, the general program is designed. At last, the improved Artificial Life Algorithm is used to solve the problem of system recognition, and the proof-test is provided.