The Research Of Artificial Fish School's Behavior And Congestion Factor

SI is a collective intelligence revealing by a group that consist by ordinary communes,this intelligent optimization algorithm inspired by insects in nature,it is a series of new methods to solve the complex problems by using a series of simulation studies on the natural behavior of insects.This new evolutionary computation technique has become more and more concerned by researchers.Artificial fish swarm algorithm(AFSA)is a swarm intelligence algorithm,and it is a new type of random search optimization algorithm,through the four basic natural imitation fish,including foraging behavior,cluster behavior and following behavior and random behavior to solve the optimization problem.Preliminary study shows that the algorithm has many excellent properties,such as parallelism,simplicity,fast out of local extreme,fast search speed and so on,but there are some shortcomings.In this paper,the problem of constant congestion factor leads to two problems,that is the search will make the algorithm run time is too long and the algorithm to get the optimal solution and the error of the actual optimal value.According to the following behavior,mechanism of poly crowded degree factor of group,the new fitness function is proposed,and the exponential attenuation variation strategy is adopted,the crowding factor increases with the number of iterations.After global search to near global extreme point precise search,It not only speeds up the convergence rate of the solution of the system,but also makes the numerical solution more stable.At the same time,the algorithm can not fast convergence,and easy to fall into local optimal problem.In this paper,we introduce the symmetric normal random behavior,we use the normal distribution to adjust the step size in the behavior,reduce the useless computation and improve the search efficiency.In addition,inspired by the artificial bee colony algorithm,the artificial fish swarm algorithm based on the following behavior is proposed.Through the observation of all the leading fish fitness value,and on the basis of the selection probability automatically determine the size of follow one lead fish,reduce the parameters of the basic artificial fish swarm algorithm in the selection,avoid the influence of non adaptive parameter settings for optimization by experience.Finally,the artificial fish swarm algorithm is applied to the three grade optimization model of antisubmarine combat mission planning,and make empirical analysis and Realization of solving the DAD model.