Multi - Population Research And Application Of Hybrid Genetic Algorithm Annealing Greed

The genetic algorithm is a kind of searching method which simulates the natural evolution. It is simple and easy to implement, especially it doesn't need the special field knowledge, so it has been used in every broad fields. Now the genetic algorithm has got a lot of fruits and more scholars begin to pay attention to it.The genetic algorithm is still a new developing technology. Despite its success in so many domains, its theoretical fundament is relatively weak. There are still lots of problems to be studied and improved.Firstly, this paper has done some work in the research and application of the genetic algorithm. A new hybrid genetic algorithm is derived, that is the multi-group annealing greedy hybrid genetic algorithm. Greedy algorithm which has more local search ability is imported to this algorithm and is well combined with the ideas of simulated annealing and multi-group parallel evolution in this paper. Simulation results show that this method not only avoids the premature convergence problem existed in genetic algorithms, but also enhances the globe convergence, thus improves the convergence velocity.Secondly, the algorithm that is improved is presented for TSP problem in this paper. By using the improved algorithm, this paper recalculates the shortest route connecting all the capitals of the 31 provinces which is 500 kilometers shorter than the widely acknowledged one. The results of the experiment demonstrate that the method is valid and efficient.Finally, the present author points out the weak points, limitation of this thesis and makes a plan for further study.