Invasive Weed Algorithm And Its Application For Shop Scheduling Problem

Job shop scheduling problem(JSP)is an important issue in the manufacturing production scheduling problem.It has a wide range of applications,mainly related to carrier dispatching,airport aircraft dispatching,port terminal cargo dispatching,automobile assembly line scheduling and other practical scheduling problems.Therefore,the study of workshop scheduling has a very important significance.Aiming at the problem that traditional mathematical methods are not meet the demand of production scheduling,a new meta-heuristic optimization algorithm is mainly studied to solve job shop scheduling problem in the thesis.Invasive weed optimization(IWO)is a new meta-heuristic algorithm inspired by weed invasion behavior.The invasive weed algorithm simulates the weed behavior of weeds and the IWO has the identical characteristics as weeds,such as strong robustness,adaptability and randomness.In the paper,the mechanism of IWO algorithm is deeply studied.The algorithm flow is analyzed to improve the parameter part of the algorithm and balance the capability of exploration and exploitation.Moreover,the IWO algorithm is applied to JSP problem.The main research contents are as follows.1.The invasive weed algorithm(IWO)is demonstrated the robustness on adaptability and efficiency in solving optimization problems.However,IWO also has the disadvantages of being significantly affected by parameters and easily falling into local optimum.In the paper,a hybrid algorithm based on adaptive intrusive weed algorithm(IWO)and differential evolution algorithm(DE),named SIWODE,is proposed to solve the single objective optimization problem.First,the two parameters in the SIWODE are adaptively operated to improve the convergence speed of the proposed algorithm.Second,the crossover and mutation operations are introduced into SIWODE to improve the diversity of population and increase the capability of exploration during the iterative process.Furthermore,a local perturbation strategy is applied to improve the exploitation during the late process.The experiment results of SIWODE show that the SIWODE has superior searching quality and stability than other mentioned approaches.2.A discrete invasive weed optimization algorithm for job shop scheduling problem(DIWO)is proposed to solve the job shop scheduling problem.As a random numerical optimization algorithm,a certain number of candidate solutions are generated in space and the candidates are encoded and decoded to operate in the algorithm.And the new individuals are generated through the propagation and diffusion in DIWO.The generation of new individuals is controlled to approach the optimal solution.Afterwards,the local search strategy and the variable neighborhood search operation(VNS)are applied to perform a local optimization and adjustment.Furthermore,the performance of the algorithm is improved and enhanced the ability of the local search and the deep search ability of the algorithm.3.The Markov model is applied to analyze the convergence of the SIWODE algorithm.The SIWODE algorithm is testified on the CEC2017 benchmark set and the DIWO algorithm is detected on the LA test set for simulation experiments.In the paper,the statistical results of the simulation experiments are statistically analyzed by the method of hypothesis testing.