Improvement And Application Of Bat Algorithm

Swarm intelligence algorithm has been an emerging evolutionary computation technology developed in recent decades.It has the characteristics of easy to implement and suitable for parallel processing.Typical algorithms,such as particle swarm optimization and ant colony algorithm,have become powerful methods for solving many difficult optimization problems.As a new kind of swarm intelligence algorithm,bat algorithm is inspired by the bat's echolocation model to search an optimization solution.The algorithm has the advantages of simple model,fast convergence,potential parallelism and distributed characteristics.Because of the super bat,on the one hand,BA can converge quickly;on the other hand,it is easy to fall into local optimum,premature convergence,slow convergence in the later.To solve the above problems of the bat algorithm,we improve the performance of the algorithm from the population cooperation and search process.We propose two improved bat algorithms:(1)Bat algorithm based on dynamic learning of opposite population.This new algorithm generates opposite population by opposite strategy,and more competitive individuals are joined in the base population by opposite learning.We also present elite choices for the current and the opposite population.In this way,the search diversity and search intensity can be enhanced,and avoid the algorithm falling in to local optimum and late stagnation;(2)Bat Algorithm based on Cross Boundary Learning and Uniform Explosion Strategy.Cross boundary learning improving the opposite learning and expanding the learning area,maintain the ability of global exploration in the process of fast convergence.In order to enhance the ability of local exploitation of the proposed algorithm,we proposed uniform explosion strategy,which can achieve almost same search precise as that of firework explosion algorithm but consume less computation resource.In order to verify the effect of the algorithm in practical problems,we apply it to the 0-1 knapsack problem by changing the coding strategy.At the same time,we combine the bat algorithm with the FCM and KFCM algorithms to apply it to clustering problems and image segmentation.