Research And Application Of New Algorithms For Pattern Matching Based On Game Theory And Bionic

Traditional pattern matching algorithms in the field of pattern recognition is that the sample pattern matches with the patterns in the pattern library actively. There are two shortcomings:First, the information of the patterns in the pattern library is not used fully, and the match results lack explanations. Second, the matching process is complicated when the pattern library is large. Facing the two shortcomings, the paper advances two kinds of new methods for pattern matching with game theory and bionic theory. Then the paper gives the core technology of the algorithms, and applies which to the defect detection problem for piezoelectric quartz crystal.In the first place, the paper advances a new method for pattern matching with game theory, which uses the information of the patterns fully and the patterns match with the sample pattern actively by the non-cooperative and cooperation game theory stages. And then comparing with the matching results by fuzzy identification, the matching results by game theory are right.Finally, the paper advances a new method for pattern matching with bionic theory, which is based on bionic high-dimensional space and random memory taboo search theory. And the method improves the matching speed, reduces the complexity of the matching process. More importantly, the confidence level of the results by bionic matching is very high.