Firely Algorithm Based PID Control Method For Overhead Crane

Firefly algorithm(FA)is a kind of swarm intelligent optimization algorithm,being simple,distributed and efficient,which has been applied to many optimization problems.FA is inspired by the wild firefly communication mechanics in the night by flashing light.Overhead cranes are very commonly used for the transportation of material.The anti?sway control problem of overhead cranes based on PID controller and the parameters optimization of PID controller using modified FA are discussed in the thesis.The main contents of the thesis are as follows.(1)Owing to the basic FA lack of adaption for solution space in different scales and shapes,a modified FA,Location Mean FA(LM-FA)with firefly group location mean value constraint is proposed.In addition,random variable step factor replaces the original static one to promote the convergence precision.The light intensity and light difference are also included in the flying formula to increase the firefly diversity preventing from early maturity.Simulation results of the test function results show that the proposed algorithm has better performance.The two PID controllers are for position and angle respectively,which is not good enough to meet the request.The parallel double PID controllers are proposed for position.Owing to the poor performance of traditional methods to optimize the parameters of PID controllers,LM-FA is used in simulation experiment based on the 2-D overhead crane mechanical model.The experiment results validate the effectiveness of the parallel double PID controllers and the LM-FA.(2)A modified FA,Locally Search Firefly Algorithm(LS-FA)based on the fitness value with local search space is proposed.The fitness value distribution is scattered in global space.As a result,the local flying based on fitness value has a reasonable search ability.In order to avoid the early maturity,a method to monitor the state of the algorithm convergence is proposed based on the fireflies distance matrix and the object function value of the firefly group.Through some typical test functions,the results show that LS-FA performs better than the basic FA.The experiments are carried on based on an experiment platform.The LS-FA is used to search the proposed PID controller parameters by off-line way.The results show that parallel double PID controllers are much better than the traditional PID controller with smaller swing angle,faster speed and without overshoot.