Study On Artificial Immune Network Models For Optimization Algorithm

As a parallel,distributed and self-adaptive information processing system with high intelligence performance, The nature immune system provides a new way to deal with actual calculation problems. Because The artificial immune system has fine characteristics on study, cognition and memory .it has great potential In network planning, problem solving combinatorial optimization.Based on Artificial immune system the principle ,Clonal selection algorithm and the immune network (aiNet) method are global optimization algorithms which be used most commonly, these two algorithms have certain advantages on searching for the optimum and avoiding falling into local minimum, but its also have some defects on keeping the diversity of solution set.In this paper, a new optimization algorithm Clonalg-aiNet be proposed for the shortcomings of these two algorithms,. The algorithm aims to find the global optimum solution while searching a local optimal solution as much as possible. This Algorithm combines the advantiages of clonal selection algorithm and the immune network algorithm, and based on the clonal selection algorithm,introduced the suppression of immune network algorithm to the evolution of each generation, selected a similar problem solved, and then according to affinity, remove the poor quality solution. The suppression remove the similarity between different populations of individuals and keep the diversity of set more better. In addition, the new algorithm improved in the selection strategy of clonal selection algorithm, which is to ensure the optimal solution can be inheritance and retention, while maintaining the population size.In this paper, the new algorithm is analyzed and simulated,and do the compartion of the results in the implementation process and performance of the traditional clonal selection algorithm and aiNet algorithm and new algorithm . Experiments show that the algorithm has more effectively insearching the global optimal solution and the local optimal solution.