Research On Group Batch Scheduling Considering Deterioration Or Learning Effect

With the continuous integration of new generation information technology and manufacturing industry,the manufacturing mode and production organization methord of the enterprises have deeply changed.Meanwhile,the production efficiency of manufacturing industry has been significantly improved.The traditional manufacturing mode for mass production and few varieties has been difficult to meet the growing personalized consumption demand.In recent years,more and more enterprises have invested and introduced intelligent scheduling systems based on group manufacturing and batch production mode.Morever,the intelligent scheduling system adopted more flexible dynamic resource model and method for precise production scheduling,so as to better adapt to the production characteristics of multi variety,small batch and personalized customization.Among the dynamic resource model,deterioration or learning effect of production resource is considered as one of the most significant characteristics.In recent years,the problem of group batch scheduling considering deterioration or learning effect of production resources has gradually become a hot topic in the research field of production scheduling.In this dissertation,based on the actual production process of discrete manufacturing industry such as auto parts and semiconductors,we systematically analyzed the multiple group batch scheduling problems in the situation of single customer and multi customers.In these studies,based on group manufacturing and batch processing mode,we considered the deterioration or learning effect of manufacturing resources,which existed in the new generation of intelligent manufacturing systems.Due to the actual production characteristics,the deterioration effect of the machines is obvious in the production process with single customer,while the deterioration effect of the machines(old factory)or the learning effect of the workers(new factory)is obvious in the production process with multi customers.Based on these production characteristics,the paper focuses on four aspects of group batch scheduling.They are about considering deterioration effect in the case of single customer,considering deterioration effect in the case of single customer with outsourcing,considering deterioration effect in the case of multiple customers and considering learning effect in the case of multiple customers.In this paper,we focus on the deep analysis of these complex scheduling problems.We abstract efficient and available scheduling models and design heuristic scheduling rules or intelligent scheduling algorithms to solve the problems.The main research results and innovations are as follows:(1)For the problem of group batch scheduling considering deteriorating effect in the case of single customer,the optimal scheduling rules for single machine and multi machines are constructed.For the multi machines scheduling problem,a new hybrid AIS-VNS algorithm is designed by using the proposed optimal structure attribute and batch processing rules.This algorithm combines the advantages of AIS and VNS.Experimental results show that the algorithm has a good advantage in efficiency and quality compared with the traditional algorithms.(2)For the problem of group batch scheduling considering deterioration effect in the case of outsourcing,the structural variables and characteristics are constructed.Based on these structures and characteristics,an effective hybrid intelligent algorithm VNS-NKEA is proposed to solve this type of problems.This algorithm can effectively solve the problem by learning from the advantages of neighborhood based evolutionary algorithm(NKEA)and variable neighborhood search algorithm(VNS).The experimental results show that the hybrid VNS-NKEA algorithm can solve this type of problems more efficiently and effectively.(3)For the problem of group batch scheduling considering deterioration effect in resource competition environment,the key structural properties and optimization models are constructed,and a decision flow chart embedded with relevant scheduling rules is designed based on these structural properties and models.Morever,an effective improved differential evolutionary(DE)search algorithm is proposed.The algorithm uses the local operation strategy of VNS algorithm for reference to solve the scheduling problem on continuous batch machines.Experimental results show that the improved DE algorithm is more effectively and stable than other similar algorithms.(4)For the problem of group batch scheduling considering learning effect in resource competition environment,the learning effect models of different job groups are constructed,and the batch scheduling rules for single machine and multi machine are proposed.For multi machines scheduling problem,an efficient iterative reference greedy algorithm(LIMA-IRG)based on “less is more” is designed.This algorithm improves the efficiency for solving this type of problems by eliminating the low efficiency steps of traditional complex algorithm.Experimental results show that LIMA-IRG algorithm has higher efficiency than other traditional algorithms.Finally,based on the core problems and the basic framework of the study,this paper also looks forward to the future research of complex scheduling problems,such as considering both deterioration effect and learning effect,considering more production characteristics and considering multi-objective optimization.