Research On Geese Swarm Flying Theory And Geese Swarm Optimization Algorithm

Formation flight of geese swarm is an optimal behavior of swarmintelligence through natural selection and evolution in the bionics, and theresearch on its intrinsic mechanism has great impact on practicaloptimization. Through in-depth analyses and researches on bionic theories offormation flight of geese swarm, the theory that geese formation flight savesmuch energy has been considered more reasonable, and new theories of theirformation flight and a new geese swarm optimization algorithm are proposedin the article. It has been proven that the new algorithm can effectively solvethese problems such as particles premature, poor search accuracy, easilytrapped into local minimal values and so on. Several aspects on researchcontents in this paper are as follows:(1) In order to solve above problems of PSO, in-depth research andimplementation on an efficient improvement to particle swarm optimization(Geese Particle Swarm Optimization, GPSO) algorithm have been achievedthrough the reference and study on geese swarm flight, and GPSO in someextent has improved optimal performances than PSO. But the defect ofunreasonable mechanism of average-weight for minimum optimization alsohas been found through experimental results.(2) For unreasonable average-weight method in GPSO, a new geese swarmoptimization algorithm based on Gaussian-weight are proposed through theimprovements of GPSO. The new algorithm can solve the problem of GPSO, testifyGaussian-weight method reasonable and effective, and improve most optimalperformances for GPSO algorithm.(3) In the foundation of research on geese flight and previous achievements, fivegeese-flight rules and hypotheses are presented to form a concise and perfectgeese-flight theory framework in this paper, and these theories are also accordancewith five basic principles of swarm intelligence. A new algorithm of geese swarmoptimization is achieved based on standard particle swam optimization algorithm. Compared with the standard particle swarm optimization algorithm and GPSOalgorithm, experimental results show that the new algorithm is advanced onconvergence, searching ability, robustness, and etc. And the results can verify thesetheories highly rational and correct.