Research On Many-objective Green Flexible Job Shop Scheduling Problem

Manufacturing industry has always been an industry with intensive energy consumption and serious environmental pollution.With the rise of "green manufacturing",the manufacturing industry is facing new challenge for green transformation.At the same time,the demand of customers for product diversification and customization also urges manufacturing enterprises to adjust their production mode to small batch and customization.Shop scheduling,especially flexible job-shop scheduling,is the key to solving the above problems,because production workshop is the main production place of manufacturing industry.At present,in the flexible job shop scheduling problem(FJSP),most scholars take the economic benefit index as the optimization goal,and pay limited attention to the green goal,and these studies often focus on the energy consumption or carbon emission,ignoring the noise,waste emission and other environmental pollution problems in the actual workshop production.In addition,decision makers usually consider production efficiency,resource utilization,production cost and other optimization objectives,which makes the solution of FJSP more difficult.Therefore,this research focuses on the mathematical model and optimization algorithm of the many-objective green flexible job-shop scheduling problem(Ma OGFJSP).Firstly,on the basis of the four traditional optimization goals of maximum completion time,production cost,customer satisfaction,and maximum machine load,this paper uses noise,energy consumption,and waste emissions to comprehensively evaluate the green degree of workshop production,and build a mixed integer programming model of Ma OGFJSP,which optimizes the traditional goals and green indicator at the same time.In this model,the processing speed,which is the processing parameter of the machine,is introduced to discuss its influence on the optimization objective.Secondly,this research uses the Reference-point based Non-dominated Sorting Genetic Algorithm(NSGA-III)to solve Ma OGFJSP.According to the characteristics of the Ma OGFJSP model,we first design a three-layer chromosome coding and decoding method,next design three kinds of population initialization strategies,after that design the crossover and mutation method suitable for Ma OGFJSP,then improve the elite strategy to aviod premature convergence or falling into local optimum,finally apply the variable neighborhood search(VNS)with two neighborhood structures to improve its local search ability.Finally,in the simulation experiment,this paper designs a test case suitable for Ma OGFJSP based on the classic case to test the overall and local performance of the hybrid algorithm,and evaluates the non dominating solution set by using the index of inverse generation distance and hypervolume.The experimental results show that the hybrid algorithm is feasible and effective in solving many-objective green scheduling problems.At the same time,taking the production line of the cooling system in the engine shop of company A as an example,this study selects a satisfactory scheduling scheme for the decision-maker through the weight of each objective function,and uses the green scheduling strategy of delayed processing strategy to further optimize the green index.In terms of theory,this research adds the knowledge of Ma OGFJSP mathematical model and optimization algorithm.From a practical point of view,this research can guide manufacturing companies to carry out green job shop scheduling,improve workshop economic efficiency,reduce resource waste and environmental impact,and help decision-makers select satisfactory solutions from the Pareto solution set.