Grasshopper Optimisation Algorithm And Its Application Research

With the development of society,the optimization problems appear more complex in many fields such as engineering application and energy management.The traditional optimization methods dealing with these problems have become increasingly difficult.The swarm intelligence optimization algorithm with bionic behavior has been using widely in various fields because of its advantages of simple implementation,high efficiency and strong robustness.The grasshopper optimization algorithm is a new swarm intelligence optimization algorithm in recent years and it has been successfully applied in many practical fields.But the grasshopper optimization algorithm has the disadvantages of easily falling into local optimum and slow convergence.Therefore,aiming at solving problems of grasshopper optimization algorithm so as to improve the performance of the algorithm and apply it in data clustering,this thesis mainly proposes the following effective improved algorithms:(1)This thesis proposes grasshopper optimization algorithm on the foundation of differential evolution and opposition learning.It performs differential evolution between grasshoppers and those that cannot find better position in each iteration of the algorithm so as to help increase the probability of jumping out of the current inferior position and looking for a better position.Thus,the grasshopper can enhance the algorithm's global search capabilities through opposition-based learning.(2)This thesis also proposes grasshopper optimization algorithm on the base of curve adaptive and simulated annealing.It improved the key parameter c of algorithm,and then,by applying the simulated annealing algorithm,which has a strong local searching ability and the ability to jump out of local optimum,to enhance the local searching ability of the grasshopper algorithm,so as to avoid the algorithm falling into local optimum.(3)To enhance the local searching ability of(1),this thesis also proposes grasshopper optimization algorithm based on double-differential evolution and opposition-based learning.This method divides the grasshopper population into two groups according to whether or not the grasshopper can find a better position in each iteration,and uses different differential evolution to each group.If the grasshopper on opposite position is no better than the current position,a more detailed local search would be performed on the grasshopper at the current position.In this thesis,the first two improved methods has been tested with the application of function optimization.The last improved method has been experimentally tested in function optimization and data clustering,and has been compared with other swarm intelligent optimization algorithms.All improved methods has been undergone tests with 10 different test functions and 5 date sets have been applied in the application of clustering.The improved algorithm compared the average,variance,minimum and maximum values among 30 consecutive experimental results.The clustering application uses three clustering indicators: F-measure,normlized mutual information and rand index.The experimental results show that the improved algorithm and its application can achieve better results.