Research On Self-adaptive Path Planning Algorithm Based On Genetic And Ant Colony Algorithms

Intelligent path finding problem stems from the field of geometry. Withthe gradual in-depth study in the areas of robot path finding, road navigation, virtualreality and computer game, intelligent path finding problem comes to an importantresearch issue. The research on artificial intelligence methods is particularly importantbecause of its global search, information feedback and self-learning etc. Bionicsalgorithms are that people gain inspiration from biological evolution and daily behavior ofbiological to solve complicated problems. Including genetic algorithms, neural networks,ant colony algorithm and other well-known algorithms.First, describes several commonly used current path planning algorithm, and comparetheir advantages and disadvantages. Focuses on artificial intelligent path planningalgorithms, genetic algorithm and ant colony algorithm. Genetic algorithm has a strongrapid global search capability, robustness, but can not use the feedback information in thesystem, easy to do a lot of unnecessary redundant computing at the latter part of algorithm,leading to decreased convergence rate. Ant colony algorithm has a good ability ofinformation feedback, but solve slow due to lack of initial pheromone. Presents a dynamicself-adaptive fusion algorithm based on two methods. It use genetic algorithm to generatea solution as the initial pheromone distribution in the early, then self-adaptive parameterfuses genetic operators dynamically in the latter part of ant colony algorithm. Not onlyimprove the convergence rate of solution, and improved ant colony algorithm convergestoo quickly to lead into the global optimum situation, enhance the global search ability ofentire algorithm.Through building the path planning environment by grid method and doingsimulation experiments under a variety of environment, obtained more optimized methodof parameter combinations by studying on the influence and relationship of parameters inevery stage of the self-adaptive fusion algorithm. At the same time, compared to the basicfusion algorithm and the proposed self-adaptive fusion algorithm, obtained a moresatisfactory result in the aspect of algorithm performance and effectiveness.