The Inter Generation Statistical Character Self-feedback Improved Genetic Algorithm Based on SVM Modeling

This paper analyzed the reasons resulting in prematurity in the genetic algorithm running procedure and put forth the concept of inter generations hamming distance, which can well reflect the running procedure universal trend and dynamic property. First, the inter generations hamming distance model was build by employing support vector machine; second, the optimization strategy was improved based on the modeling results and the dynamic change of the feature. By dynamically adjusting the population diversity according to the running condition of algorithm, prematurity of genetic algorithms can be effectively avoided. The numeric test results showed that the search integrity of improving algorithm had been enhanced, the search efficiency was better than that of standard genetic algorithms and the algorithm could improve the global optimization handling capacity.