The Research And Application Of Dual-Population Ant Colony Optimization Based On Dynamic Feedback Mechanism

With the progress of science and technology,the development of mobile robots is more and more rapid,which brings great changes to people's clothing,food,housing and transportation.It also belongs to a very important branch and direction of artificial intelligence and one of the most popular research directions at present.The most important modules of mobile robot are machine vision technology,Slam and robot path planning technology,which tell the robot where the obstacles,the current spatial location and how to choose the path to the destination.At present,there are many methods for solving mobile robot path planning,which can be roughly divided into four categories: graphics methods,traditional robot algorithms,intelligent bionic optimization algorithms,and other algorithms.Therefore,in order to solve the important problem of robot path planning,this paper proposes a dual population ant colony optimization based on dynamic feedback mechanism.It conducts a series of research and simulations on traveling salesman problem and path planning,and finally applies the algorithm to robots.The main research directions and contents of this article are as follows:First of all,in order to solve the influence of parameters in the algorithm on the experimental results,Particle Swarm Optimization(PSO)is introduced to optimize multiple parameters of ant colony optimization in three-dimensional space to improve the quality of understanding;at the same time,a new way of path construction is introduced: unit-distance pheromone path construction operator,to strengthen the synergy of distance factor and pheromone factor.The two path building methods communicate with each other to find the global optimal solution.In the TSP problem,through the comparison and analysis with other kinds of optimization algorithms,it shows that this algorithm enhances the population diversity of the algorithm and enhances the quality of understanding.Secondly,in order to strengthen learning and competition among populations,two classic ant colony optimization-Ant Colony System(ACS)and MAX-MIN Ant System(MMAS)are combined.The diversity of MMAS is retained,and the advantage of fast convergence speed of ACS is retained,which jointly promotes the algorithm to find the optimal solution on large-scale TSPs.The idea of collaborative filtering in the recommendation algorithm is introduced to reward the more preferred paths of ants in the two populations,making the algorithm more oriented and accelerating the algorithm's convergence speed.At the same time,the two populations are adaptively adjusted based on the dynamic feedback of the information.The frequency of communication increases the diversity of the algorithm.When the algorithm stagnates,the two populations interact cooperatively,homogenizing the pheromone of each population,and jumping out of the local optimum.In the simulation experiments of the medium and large-scale TSPs,the algorithm in this paper improves the quality of understanding,guarantees the diversity of the algorithm,and accelerates the convergence speed of the algorithm.Next,in order to verify the application of the algorithm on a larger scale,the population of the algorithm is expanded to three,and the unit-distance pheromone path construction operator is taken as a population to form multiple populations with ACS and MMAS.The Pearson correlation coefficient is used as an evaluation standard to measure the similarity between populations.Two populations with high similarity are selected and the parameters of similar paths in the two populations are rewarded to speed up the algorithm's convergence speed.According to the information between the populations,adaptively adjust the communication frequency of the two populations,increasing the diversity of the algorithm;when the algorithm stagnates,initialize the sum of the reward parameters and jump out of the local optimum.The experimental results show that the heterogeneous multiple colony ant colony optimization proposed in this paper has better experimental results on TSP than the traditional single population ant colony optimization and other multiple colony ant colony optimization,and on the large-scale problem,this advantage more obvious.Finally,with the support of the theoretical analysis,using machine vision and radar technology to collect maps in environment,and using the algorithm proposed in this paper to simulate and analyze the robot path planning.Experimental results show that the improved algorithm has good feasibility and practicability both in simulation and in practice.