Improvement Of Grey Wolf Optimizer And Its Application In Robot Path Planning

With the progress of technology and the development of society,the difficulty of solving various optimization problems is also increasing,and the traditional algorithms have been difficult to meet people’s increasing computing needs.In recent years,researchers have found that heuristic algorithm has unique advantages in solving optimization problems.It is one of the very effective methods and technologies to solve optimization problems.It can find the high-quality solution of this kind of problem under multiple constraints.Grey wolf optimizer adopts adaptive convergence factor and optimal leader diversity strategy.It has the characteristics of clear structure and few configuration parameters,and has good optimization performance.However,there are such as unstable optimization results and difficult to jump out of local extremum and low solution accuracy,which need to be further improved.Focusing on the shortcomings of grey wolf optimizer paper deeply studies and analyzes the search mechanism of the algorithm,and puts forward a new improved algorithm.The improved algorithm is used to solve CEC test functions,engineering constraint optimization problems and robot path planning problems.The main work of this paper is summarized as follows:(1)An improved grey wolf optimizer(FMGWO)which integrates FPA,teaching mechanism and multinomial mutation is proposed.Firstly,the GWO and FPA are combined.The prey search stage of grey wolf is improved by Levy flight mechanism,and the local search stage of grey wolf optimizer is improved by the double random mechanism.The optimization ability of the algorithm is enhanced.Then,the teaching mechanism is added to the α wolf,accelerate the approach of grey wolf’s individuals to the optimal value region;Finally,the individuals with poor optimization effect are conducted position variation to increase the diversity of the population.In order to verify the operation efficiency of FMGWO,the time complexity of GWO and FMGWO algorithm is analyzed.The time complexity of the two algorithms is the same.The test results of five algorithms on different dimensions of CEC2017 benchmark functions show that FMGWO algorithm has high solution ability.(2)An improved grey wolf optimizer(GGWO)based on crossover strategy,adaptive adjustment of convergence factor and t-distribution disturbance mutation is proposed.Firstly,the horizontal crossover operation is carried out for two different individuals in the population in all dimensions,so as to reduce the search blind spots,and the vertical crossover operation is carried out for the two different dimensions of the optimal individual,The ability of the algorithm to jump out of the local optimum is increased.Then,the convergence factor is adjusted nonlinearly according to the number of iterations.Finally,the t-distribution perturbation mutation strategy with the number of iterations as the degree of freedom is added to improve the accuracy of the algorithm.By comparing the improved algorithm with five representative comparison algorithms,the tests on CEC2017 benchmark functions,more challenging CEC2021 benchmark functions and four engineering optimization design problems show that CGWO algorithm has obvious advantages,and has better solution ability and application potential.(3)FMGWO algorithm is applied to solve the robot path planning problem in 2D circumstance,and the environment modeling is combined with the path planning algorithm for optimization.The grid method is used for environmental modeling,the environmental models with different complexity are constructed,the individual coding way suitable for solving the question is ascertained,and the fitness function,safety obstacle avoidance and line smoothing methods are constructed.The simulation experiments show that FMGWO algorithm has great advantages for figuring out robot path planning in 2D environment.Based on the research of 2D robot path planning,the 2D grid method is extended to 3D space,and then the CGWO algorithm is applied to solve the robot path planning problem in 3D environment.The simulation results show that CGWO algorithm is obviously superior to other comparative algorithms in solving the robot path planning problem in 3D environment.