Joint Optimization Algorithms For Minimizing Total Tardiness In Seru Production

With the promotion of global economy,more and more advanced techniques appears,adding unprecedented chances and challenges to the manufacturing industries in China.In this important period,it is necessary for Chinese manufacturing companies to study foreign advanced production modes and techniques.Therefore,this paper focuses on a new production mode,Seru Production,which derives from Japan,to carry out investigation.Seru is a highly self-finishing assembly unit,made up with several multi-skilled workers.A seru can finish most part,even all of the assembly tasks of a piece of product.Seru Production overcomes many drawbacks existing in assembly line production,and possesses unique advantages in the fluctuating market,such as fast response capability.Minimizing total tardiness is crucial to enhancing customer satisfaction and competitiveness for a company,so the objective has attracted wide attention in the optimization field of manufacturing organization.However,there has not been any research on minimizing total tardiness in Seru Production,thus we deeply investigated the problem.The operation of seru system contains two main decision processes,Seru Formation(SF)and Seru Scheduling(SS),which have been both proved NP-hard.To simplify the problem,the previous investigations only optimized one of decision processes with the other given.Solving seru formation problem with seru scheduling following a certain scheduling rule is a common method.However,such methods cannot obtain the global optimal solution.Therefore,this paper focuses on exploiting joint optimization algorithms for minimizing total tardiness in seru production.In this paper we:(1)Build the model for minimizing total tardiness in Seru Production.Based on the operation model for minimizing makespan in seru production,we build another model for minimizing total tardiness.Besides,we analyze the complexity of joint optimization,drawing the conclusion that the solution space expands exponentially with the increase of problem scale.(2)Exploit two exact algorithms.One is the exact algorithm for minimizing total tardiness in seru production basing on backtracking algorithm:First,enumerate all the seru formations with recursion algorithm.Then,for each seru formation,obtain the optimal seru scheduling with backtracking algorithm.Finally,obtain the optimal solution through comparing the objective values of the optimal seru schedulings of all the formations.The other algorithm bases on the linear mixed integer model of job shop scheduling:Build the linear seru scheduling model referring to the linear mixed integer model.For each seru formation,obtain the optimal seru scheduling using CPLEX,a linear optimization tool.The two proposed exact algorithms are proved to obtain global optimal solutions of small scale cases.(3)Exploit an intelligent algorithm basing on cooperation coevolution mechanism:adopt Genetic Algorithm(GA)to deal with seru formation and Anti Colony Algorithm to deal with seru scheduling respectively.When seru formation evolves,the best seru scheduling assists to evaluate the individuals.The best seru formation will update after the evolution and then assist seru scheduling to evolve.Then the best seru scheduling updates and assists seru formation to evolve.The circulation repeats until the algorithm comes to convergence.The intelligent algorithm is proved to obtain the global optimal solutions of small scale cases and satisfying solutions of large cases in a reasonable time span.Compared with assembly line,the proposed joint optimization algorithms of seru production sharply decreases the total tardiness.Finally,we put forward several proposals for total tardiness minimization in seru production,which have reference value for manufacturing companies to study and implement seru production.