Font Size: a A A

Multi-objective Optimization Algorithm And Its Application In Path Planning Of Mobile Robot

Posted on:2019-02-06Degree:MasterType:Thesis
Country:ChinaCandidate:H Y WangFull Text:PDF
GTID:2348330569978164Subject:Control theory and control engineering
Abstract/Summary:PDF Full Text Request
The multi-objective optimization problem is ubiquitous in practical applications.It is often used in areas such as engineering design,process design,investment portfolio,shop scheduling,grid configuration,rail transit,and logistics routes.Multi-objective optimization needs to consider the optimality of more than two goals at the same time.In practice,there are often complex relationships among multiple objectives.One goal improvement may lead to losses on other objectives.The decision makers hope to achieve each of them.It is often difficult to make trade-offs among multiple objectives on the goal of optimality,so it is of great importance to study multi-objective optimization algorithms.In this thesis,the NSGA-II(Nondominated Sorting Genetic Algorithms II)algorithm and the RPEA(Reference Points-based Evolutionary Algorithm)algorithm based on the reference point based optimization algorithm are introduced in detail,and some improvements are made on the basis of the NSGA-II algorithm:1)An elite replacement strategy is proposed to speed up the survival of the fittest.Each evolutionary generation of the population selects several optimal solutions from the previous population to replace the worst performing solutions in the current population.2)Analyzing and pointing out the deficiency of adaptive genetic algorithm in adjusting genetic parameters,and improve the adjustment strategy of genetic parameters in light of its deficiency: adjusting the crossover probability in stages and adjusting the mutation probability adaptively.The standard test function set ZDT and DTLZ are used to test the improved algorithm and compare with the original NSGA-II algorithm.The experimental results show that the improved algorithm is superior to the original NSGA-II algorithm in distribution and convergence.The reference point-based evolutionary algorithm RPEA adopts a new generation reference point strategy: First,the non-dominated optimal solution in the population is obtained,and a smaller value is subtracted from each sub-objective of the non-dominated optimal solution to obtain a candidate reference point.Secondly,the individuals who are close to the reference point in the population are selected to the next generation.Through the continuous evolution of the population,the reference point gradually approaches the true Pareto front of the problem,and the Pareto optimal solution of the multi-objective optimization problem is obtained.Finally,through the experimental analysis of the basic test functions ZDT and DTLZ,the performance of the RPEA algorithm and the improved NSGA-II algorithm in solving multi-objective optimization problems are compared.The result verifies that the RPEA algorithm is superior to the improved NSGA-II algorithm in solving performance.The above two multi-objective optimization algorithms are applied to the path planning of mobile robot.Firstly,the coding methods,crossover operators and mutation operators adopted in genetic operations are introduced.And improve the path selection method.Then,the mobile robot's movement environment model is built,and the path planning of mobile robot is realized by using Matlab software.Finally,through experimental comparison and analysis,it is verified that the path obtained by the improved NSGA-II algorithm is superior to the path obtained by the traditional NSGA-II algorithm both in path length and safety,and the running time of the improved NSGA-II algorithm is shorter.The simulation results using RPEA algorithm to solve the path planning problem verify that the RPEA algorithm is feasible in solving the path planning problem of mobile robot.
Keywords/Search Tags:Multi-objective optimization, NSGA-? algorithm, reference point-based evolutionary algorithm RPEA, crossover mutation probability, path planning for mobile robot
PDF Full Text Request
Related items