Research And Application Of Natural Computing Based On Multiple Competitive Elimination Strategy

In recent years,in order to meet the growing demand for computing,natural computing comes into being.These heuristic algorithms have unique advantages in solving optimization problems.However,in natural computing methods,in order to solve the optimization problem of high-dimensional data,the population size needs to be increased to obtain higher accuracy,but at the same time,the time complexity is relatively large.If the population size decreases,the algorithm will fall into local optimal phenomenon.In order to solve the problems such as difficult population balance,slow algorithm convergence and easy to fall into local optimum in the optimization process,and different heuristic algorithms have different focuses,this paper studies and analyzes the search mechanism in depth for the general problem of natural computing,proposes a new general improvement strategy,and applies the improved algorithm to engineering design constraint optimization and robot path planning.The overall work content of this paper is summarized as follows:(1)A Natural calculation method based on Greedy-renewal and Self-adaption is proposed Greedy-renewal and Self-adaption Disturb(GSD)are universal strategies that can effectively improve the optimization accuracy of algorithms and have the ability to jump out of local optimal.Firstly,chaotic mapping is used to initialize the population.The quality of the initial solution is increased,and then,in the optimization iteration process of the algorithm,the greedy retention mechanism is used to carry out dimensional update on the algorithm individual to retain the solution with higher fitness and accelerate the algorithm convergence speed.Finally,according to the number of iterations,the adaptive adjustment disturbance factor is introduced to probabilistically disturb the individual to improve the diversity of the population in the later period and avoid the local optimal.The convergence performance of the algorithm is tested under different dimensions,and the results show that the strategy has better universality and optimization ability.(2)The Natural calculation method based on Multiple Competition Elimination(MCE)is proposed,which is a universal strategy.It can effectively solve the problem of low optimization accuracy and easy to fall into local optimal in high dimensional problems.First,the original solution space was divided into two types of large space with competitive relationship,and each type of large space was decomposed into N-yuan small space.Then,different elimination methods were carried out in the two types of large space to eliminate the poor individuals.Finally,some better individuals in N-yuan small space were selected to carry out competitive exchange across the two types of large space to maintain the diversity of the whole population.Thus,the convergence speed and accuracy of the algorithm are improved.The theoretical analysis and experiments show that the multiple competition elimination(MCE)algorithm can effectively improve the algorithm performance without increasing the time complexity.The experimental results show that the solving ability of the improved MCE algorithm is superior to other comparison algorithms.(3)The improved strategy is applied to natural calculation and tested in four engineering design constraint design problems.The experimental results show that the convergence performance and convergence accuracy of the improved algorithm are better than other algorithms.The improved algorithm of MCE strategy is applied to the path planning problem of robots in a two-dimensional environment.The semi-random method improved by dijkstra algorithm is adopted to initialize the population,ensure the quality of the initial population,reduce the backtrack and other conditions affecting the convergence performance.The grid method is used for modeling,the fitness function is constructed,and the path smoothing processing mode is designed.The simulation results show that the improved algorithm of MCE strategy and GSD strategy has better optimization performance than the comparison algorithm in solving the robot path planning problem.