The Study On Improvement Of Shuffled Frog Leaping Algorithm

Intelligent computation is a way to learn the natural world which is inspired by the principle of nature to imitate biological behavior for designing algorithm.Therefore,intelligent computation is also known as "soft computing".The charming characteristic of the intelligent optimization algorithm lies in the fact solving complex problems because of simple principles and model realization of this algorithm.Compared with the classical algorithm,the intelligent optimization algorithm has significant improvement when dealing with some complicated optimization problems such as problems with multiple variables,conditions,dovi,objects and peaks.The Shuffled Frog Leaping Algorithm has better practicability,adaption,stability and efficiency.Shuffled Frog Leaping Algorithm(SFLA)is a new kind of swarm intelligence optimization algorithm according to the frog behaviors.This algorithm not only has the advantages of the local heuristic search of meme algorithm,but also has the advantages of the global search of particle swarm optimization algorithm.The leapfrog algorithm is an effective optimization algorithm which owns many advantages such as its powerful search ability,simple concepts,strong robust,higher speed of convergence and fewer involved parameters,and easy realization,therefore it has been widely used in many fields such as communication technology,signal processing,image processing,the distribution of water resources and control theory.Leapfrog algorithm is applied to solve practical optimization problem more frequently,and has achieved very good results.However,the leapfrog algorithm is relatively young and the theoretical foundation is still weak.Unluckily and naturally,there are still some deficits about this algorithm,for example,it is easy to fall into local extremum and the convergence precision is not high enough.Thus we expect all of these would be developed and improved in the future.Considering the limitation of the leapfrog algorithm in the application to constrained optimization problems,we make progress on the aspects of low search accuracy and its tendency to local optimum.The main research includes the following two points:(1)This paper putting forward a Shuffled Frog Leaping Algorithm(SFLA)which isapplied to solve constrained optimization problems.Combined with the?-differential evolution algorithm(?-DE),SFLA makes full use of the information of infeasible solutions in the population during the evolutionary process.The improved SFLA not only has higher convergence speed and higher accuracy and need use less computational resource,but also solve constrained optimization problems more effectively.(2)Proposing a search strategy based on Gravity.Gravity Shuffled Frog Leaping Algorithm(GSFLA)introduces an idea for calculating inertial mass.and make full use of the information of mutual influence and co-evolutionary in the subgroup of frogs,which effectively improves the performance of the algorithm,in accuracy and convergence thus it has better practicability.The classic function simulation experiments verity the superiority of this algorithm.Finally,this paper gives a comprehensive and systematical summary and points out aspects where the SFLA need improve.