Improved Bat Algorithm And Its Applications To Robot Path Planning

With the continuous development of social economy,various optimization problems have become more and more complex.The traditional numerical calculation methods can not adapt to the rapid development of society.In order to explore new methods for solving optimization problems,researchers have been inspired by the living habits and laws of living creatures in nature,and proposed many heuristic intelligent optimization algorithms.By virtue of their own advantages,these algorithms have made due contributions to solving various optimization problems,and have been applied in many fields.In 2010,Cambridge University scholar Yang proposed a new heuristic intelligent optimization algorithm---bat algorithm(Bat Algorithm,BA).The bat algorithm is proposed by simulating the bat to emit and receive ultrasound to prey.At present,bat algorithm has been applied in interactive testing,reactor core arrangement,continuous optimization,traveling salesman,numerical optimization and so on.Although bat algorithm has been applied in many fields,bat algorithm is still easy to fall into local extremum,the accuracy of optimization is not high,and the convergence speed of the algorithm is slow at the later stage.In view of these shortcomings,this paper analyzes the reasons for these shortcomings on the basis of consulting a large number of relevant literature and continuous simulation experiments,and makes corresponding improvements.The improved algorithm is applied to the practical problems of function optimization and robot path planning.The main tasks are as follows:(1)In order to solve the optimization problem of function extremum,aiming at the shortcomings of the basic bat algorithm,such as easy to fall into local extremum,low precision of optimization,slow convergence of the algorithm in the late stage of the algorithm,and so on.A bat optimization algorithm with moderate orientation and perturbation of trend(OPBA)is proposed in this paper.Firstly,in the velocity update formula of the global search stage,a nonlinear variation factor is added to control the moving step size of the bat population,thereby improving the global exploration ability of the algorithm.At the same time,we improve the location update method of the bat in the local search stage,thereby improving the depth mining ability of the algorithm local search.Then,the trend random disturbance mechanism is introduced to further improve the convergence accuracy of the algorithm.Finally,theimproved algorithm is compared with the basic algorithm and other improved algorithms through 10 test functions,and the convergence images of the optimal solution,the worst solution,the average solution and the function are obtained respectively.The analysis of these data proves the effectiveness of the OPBA algorithm in solving the function extreme value optimization problem.(2)In order to explore a better solution to the problem of robot path planning and to expand the application field of bat algorithm,a bat algorithm with dynamic perturbation,tangent stochastic exploration and reverse learning selection mechanism(PTRBA)is proposed.First,in the global search phase of the algorithm,the dynamic disturbance coefficient is introduced to expand the diversity of the bat population.Second,in the local search stage of the algorithm,the random exploration method is improved to improve the convergence accuracy of the algorithm.Third,the reverse learning selection strategy is adopted to further balance the global exploration and local mining capabilities of the algorithm.Through the nine test functions,the improved algorithm is compared with the other three algorithms.The experimental results prove the feasibility and stability of the improved PTRBA algorithm.(3)Combining the PTRBA algorithm with the cubic spline interpolation method,the path-based coding method is defined,and the fitness function and solution method for the mobile robot to avoid the obstacle and the shortest path are constructed to solve the robot path planning problem.In the simple and complex obstacle environment,the path planning comparison experiments of single robot and multi-robot system are carried out respectively.The feasibility and effectiveness of the improved PTRBA algorithm with cubic spline interpolation method for solving the global path planning problem of the robot are verified.