Research On Multi-objective Batch Scheduling Algorithm Based On Adaptive Clustering

Scheduling problem is widely existed in reality.A feasible scheduling scheme is able to support the decision maker in manufacturing,logistics,medical,aviation,and so on.Batch scheduling,abstracted from real production process,is a type of modern scheduling problems.Since green scheduling has become one of the symbols for manufacturing modernization,research on minimization energy consumption in batch scheduling problem has important theoretical significance and practical value.A multi-objective batch scheduling problem that minimizes both makespan and total electricity cost is studied in this thesis.Aiming at the studied multi-objective batch scheduling problem,considering that evolutionary algorithms have advantages in solving multi-objective problems and adaptive clustering is suitable to extract the distribution characteristics of solutions used to guide algorithm search.Hence,a multi-objective evolutionary algorithm based on adaptive clustering is proposed to solve the studied problem.Three strategies are designed in the proposed algorithm.First,the improved adaptive clustering method is used to extract the distribution characteristics of the solutions in search space,which can be used to guide search.Second,a recombination constraint strategy is designed based on the distribution characteristics obtained by clustering during the reproduction phase,which allows individuals to mate with each other deliberately and choose mating methods through recombination mating probability,so that the algorithm has better performance in different periods.Third,due to the diverse search requirements in different evolutionary stages,the recombination mating probability is dynamically adjusted according to the historical information of the solutions,ensuring the balance between exploitation and exploration of the algorithm.Finally,to verify the performance of the proposed algorithm,simulated experiments are performed on randomly generated testing instances.The proposed algorithm is tested in two aspects.First,the designed three strategies of the algorithm are tested.Experimental results show that both the convergence and the diversity of the solutions are improved due to the designed three strategies,so that better solutions are able to be found.Second,the proposedalgorithm is compared with other algorithms,the comparative results show that the proposed algorithm is more effective and more competitive on addressing the studied multi-objective batch scheduling problem.