Improvement Of Bird Swarm Algorithm And Its Application

The idea of swarm intelligence algorithm is derived from some habits and behaviors of social organisms.Due to the high efficiency of swarm intelligence algorithm in optimization problems,it has been widely researched.Although BSA has characteristics of simple flow,fast convergence and easy implementation,it still cannot effectively solve the problem of falling into local optimality.It is necessary to research on the improvement of BSA.The main works in this dissertation are as follows:1.An improved bird swarm algorithm of multi-process searching based on aggregation.In the later stage of BSA,birds will gradually gather together,making it difficult to accurately identify the location of the birds,which will cause the diversity of bird flocks to weaken and thus fall into a local optimum.Therefore,this dissertation will define the concept of aggregation according to the flight behavior of birds,which can judge the diversity of birds as a verification index.Due to the fact that the bird swarm algorithm easily plunges into local optima and premature convergence,this dissertation depicts the change of flock position with concentration degree during foraging to increase the feasibility of birds' searching scope,and assigns optimization position in probability,so as to increase the diversity of birds.A learning mechanism is introduced to overcome the monotony of individual information communication in the flock.2.An Improved Bird Swarm Algorithm Based on two kinds of behavior strategies.The poor diversity of BSA in the late iterations will lead to the problem of premature convergence.In this dissertation,self-learning coefficient is used to adjusts the balance between the local optimum and the global optimum.Aggregation degree is also used to describe the diversity of the population.By setting the degree of aggregation,the location of bird flock after searching is alienated to increase the diversity of population,which makes the algorithm easier to jump out of local optimum.3.Application of improved bird swarm algorithm in mechanical design.On the basis of the improved producers,this dissertation introduces the perturbation strategy whichmakes the BSA jump out of the local optimum.We apply the improved algorithm to solve mechanical optimization problems and verify the performance of the algorithm by data analysis.The bird swarm algorithm is studied deeply,which makes the bird swarm algorithm more perfect and mature,and it is of great significance to apply it to practical problems.