An improve self-adaption NGA with predator

Because have very high ability of overall situation searching and convergence speed, show excellently in keeping solution variety, the niche genetic algorithm(NGA) is widely used to solving various kinds of combination optimization problem, but the traditional niche genetic algorithm(T-GA) have the problem that the discrimination standard of Euclidean distance between two individuals is not development and change with algorithm´s evolution process, and it unable to avoid the propagate of inferior solution when keep the variety of solution, so reduced the speed of convergence and operating greatly. To counter these questions, this paper has proposed an improved self-adaptation NGA with predator, in this algorithm, at first, improved the discrimination standard of Euclidean distance, make it can change with the evolution process; then introduce the concept of predator in the artificial life algorithm, settle predator to clear up and limit the propagate of inferior solution; Finally, through use this algorithm to solve the 0-1 knapsack questions, proved that the improvement of discrimination standard of Euclidean distance and the introduction of predator heighten the efficiency of algorithm greatly.