The Application of an Improved Artificial Fish Swarm Optimization Algorithm to the Real Submersible Path-planning
|School||Harbin Engineering University|
|Keywords||PSO Chaos AFSA Swallowed behavior path planning Underwater-vehicle|
In our life, there are a lot of problems to be solved by the optimization. The swarmintelligence is a kind of optimization algorithm to be often studied and used by people. Theswarm intelligence optimization algorithm is based on bionics as a model to optimize multiplesolutions or more individuals in the whole group, in a special interaction and transfer ofinformation, finishing ultimately optimization. Now particle swarm optimization（PSO）,antcolony algorithm(ACO), genetic algorithm(GA) and BP neural network algorithm is oftenstudied relatively mature optimization algorithm. As the science and technology continue toprogress and continuous development of industry, the importance of optimization algorithmgradually reflected. Attendant for optimizing the effectiveness and more accurate of thecomputational requirements. Research in recent years, the people on the intelligentoptimization algorithm for continuous improvement to speed up the convergence rate andimprove his accuracy optimization, reducing a significant impact about environmental factors,in order to better application in the practical problems.The particle swarm optimization algorithm is widely studied, its thinking is simple, theprogramming is relatively easy to implement and can be better used in practical engineeringproblems. The particle swarm algorithm is an optimization algorithm by simulating birdsforaging behavior. Now the particle swarm optimization improved algorithm itself and learnthe advantages of the algorithm and into other algorithms. The algorithm mainly concentratedin the selection of the parameter, convergence analysis, restrictions on the initial conditions,and exploring the behavior of biological activity. In terms of convergence, particle swarmoptimization algorithm with ant colony algorithm, genetic algorithm and simulated annealingalgorithm are researching appropriately.Artificial fish swarm algorithm is a new algorithm. The formation and use of relativelylate, an optimization algorithm is derived by the analog fish behavior of life. Currently, peoplefocus on the theory of the fish swarm algorithm. The implementation of the algorithm is easyrelatively and constraint condition is less, so the artificial fish algorithm has great potentialapplications in practical problems.This paper mainly completes the following three items of the work: The first job was improved particle swarm appropriate re-structure about the inertiaweight in nonlinear slow decline, so that the inertia weight better change will increase theeffect of global optimization, to accelerate the optimization speed. Then change the learningfactor, allowed to change the form with the optimization needs. In each period has a differentlearning ability, and enhance the effect of optimization. Finally, the improved particle swarmoptimization algorithm with chaos optimization algorithm fusion algorithm better to avoidfalling into local minimum value. The content in this paper is the second chapter.The second work is introduced the intelligent optimization algorithm-artificial fish-swarm algorithm. In this paper, the forming reason of artificial fish-swarm algorithm, workingprinciple, algorithm description, influence of related parameters on the quality of research,and the algorithm itself are described in detail. In the view of the problems with artificialfish-swarm algorithm is improved properly, to retain the advantages of adding algorithm’sdevour ability. Finally, the improved algorithm and the second chapter of the particle swarmalgorithm are fusing, thus forming a new improved fish-swarm algorithm-MCP-AFSA.Through the simulation test to the use of test functions, the results prove that the newalgorithm has a fast convergence speed, high search precision, can find easily a better globaloptimal value. This part is in Chapter3and in Chapter4.The third task is to apply this algorithm in the path planning of underwater vehicle.Underwater three-dimensional space to create the model, and extracting the true depth data,three-dimensional space is transformed into a two-dimensional plane. Experimental resultscan be seen, the MCP-AFSA algorithm can find a better path forward.