Application Research Of Improved Intelligent Algorithm In Path Planning

The rapid development of society benefits from the powerful promotion of science and technology,which has deeply affected and changed the clothing,food,shelter and transportation of human beings.Various complex path planning problems indoors and outdoors are constantly being refreshed and solved.Obviously,the traditional algorithms and graphics methods for solving such problems have been difficult to meet the actual and potential needs.However,the rise of intelligent bionic algorithms represented by ant colony algorithm and particle swarm algorithm provides a new idea and direction for solving such problems.The intelligent bionic algorithm only contains some basic operations of mathematics,and the calculation is relatively simple and relatively easy to implement.Due to the shortcomings of the algorithm itself,the solution effect is often unsatisfactory,which requires some appropriate improvement and optimization.After analyzing the advantages and disadvantages of the two algorithms,this paper uses the original algorithm as the basis and optimizes it to make appropriate improvements.It is also applied to test cases for verification.The main research work of this article is as follows:(1)The shortcomings of the classic ant colony algorithm in solving vehicle path planning are studied and analyzed,and an adaptive dynamic search ant colony algorithm(ADACO)is proposed.Firstly,the TSP problem in the test case is used as the basis for experimental configuration of composite parameters and algorithm model;Secondly,the strategy of combining pseudo-random distribution with adaptive transition probability is adopted to help the population choose a higher quality path.At the same time,the pheromone intensity was segmented to effectively induce the population to jump out of the local dilemma and construct a new solution.Finally,the test results show that ADACO algorithm has been significantly improved compared to other algorithms in terms of time and distribution cost,which fully verifies the feasibility of the algorithm.(2)Aiming at the shortcomings of basic particle swarm optimization on high-dimensional complex problems,such as low population diversity,easy to fall into local "misunderstanding" and poor convergence performance,a hybrid strategy particle swarm optimization algorithm(NOPSO)that introduces niche and dynamic reverse learning isproposed.One is to use niche technology and fitness sharing mechanisms to maintain the individual diversity of the initial population and enhance the "search" ability of the population.The second is to implement the dynamic reverse learning strategy on the global elite individuals to get the corresponding reverse individuals,which can effectively stimulate the spatial "development" ability of the population and ensure the global optimization ability of the algorithm.By comparing the traditional PSO algorithm and NOPSO algorithm in different raster map environments,the feasibility and reliability of NOPSO algorithm are proved,and the planned path is obviously better than the traditional PSO algorithm in terms of advantages and disadvantages and effects.