Swarm Intelligence Algorithms And Their Application On Nagivation Plan For Underwater Vehicle

Swarm intelligence which is a new kind of optimization algorithm is be been widely concern and research for the last few years, for example particle swarm optimization algorithm(PSO) and ant colony optimization(ACO) etc; all of these algorithms are the emulations that comes from to look for food to the nature living creature system. Be discovered by research, swarm intelligence algorithm have the advantage to compare with the excellent traditional type, therefore got the favor of many researchers.PSO is more excellent opposite than a traditional algorithm to have simple principle, carry out easily, less parameter and speed of convergence quickly etc; And because PSO has good portability, it has been carried on a lot of engineerings physically. The PSO research mainly includes 2 directions:The research of the algorithm and the application study of the algorithm. The research of the algorithm mainly includes proof of astringency, improving of speed of convergence and with other excellent algorithms fusion research.ACO is imitate the behavior of ant community to look for food, by adding and vaporizing the pheromones to control the direction that the ant goes forward, and finally carry out all ants to walk the same way. It has simple principle and less parameter, mainly concentration at the adding and vaporizing of pheromones. It have advantage in searching path compared with PSO, but at the same time have disadvantaged in speed of convergence. Currently, the ACO research mainly includes parameter selection, the application study of the mixture intelligence calculate way with other algorithms.This article mainly includes three works. The first work is to carry on an improvement for the parameter of PSO:(1) the improvement for inertia weight, put forward a kind of nonlinear concave function, it mainly contain hyperbolic function and exponential function, and finally carry out a kind of dynamic adjustment with an mutations factor. (2) The improvement of the speed formula put forward small community cognition method and give two kinds of structure forms. Then through simulation, compare the performance of the improved algorithm with the basic PSO. The PSO is easily sink into local optimal in the iterative late, but the modified PSO with the improvement of mutations factor added later hopping probability, which increases the search of possibilities. Experimental results show that the modified PSO has very good convergence and the convergence speed than linear variation of basic particle swarm algorithm quick 4-5 times, convergence precision better than basic particle swarm algorithm too.This part of contents is in chapter 5. The second work is the improvement research to the ACO:(1) the improvement of pheromone updating ways for ACO, proposed a pheromone update mechanisms based on PSO. (2) on the basis of (1), introducing atavism-cluster-dependent mechanisms, namely in the iteration process of the ants group formation and cluster-dependent alternates alternately, enhance the ant global searching capability. Finally, simulation experiments, using four typical Traveling salesman problem (TSP) to verify the performance of the modified ACO. The pheromone update method decisions the convergence accuracy and convergence time of the ACO. When scale expanding, the basic ACO is easy sink into local optimal, but the modified ACO can reduce a path of pheromone accumulation overmuch. It can jump out of local optimal. This part of the content in this chapter 6.The third work is navigation planning research for underwater vehicle(UV. Path planning is the research hotspots for UV, in order to get a better way, the researchers studied many methods, such as artificial field potential method, octree regional segmentation method, etc. This paper is to explore the use swarm intelligent algorithm to path planning for UV, mainly based on swarm intelligence algorithm to analyze environment modeling, translation navigation problem into the optimization problem, and give navigation algorithms and path optimization algorithm. The improved particle swarm optimization and the ant colony optimization are applied in to solve the path planning. From the experimental results we can know, swarm intelligence algorithm can be basically realized navigation planning of UV in 3d underwater. This part of the contents in this chapter 4,5 and 6.Finally, summarize and prospect the research of this paper.