Research On The Reconfiguration Optimization Method Of Personalized Customized Virtual Production Line In Fog-manufacturing

The rapid changes in diversified market demands have accelerated the upgrading of personalized products,posing a considerable challenge to the traditional enterprise’s manufacturing model of "one generation of products and one production line".It is timeconsuming and laborious to build a physical production line for short-cycle,small and medium-sized personalized orders,and the gain outweighs the loss.Therefore,there is a great demand for research on virtual reconfiguration based on multi-physical production lines.At present,the research on production line reconfiguration mainly focuses on the static reconfiguration of new physical production lines for specific types of workpiece set orders.The initial investment cost is high,the reconstruction period is long,and it is challenging to convert related theories to engineering.In addition,because the rhythm of the physical production line reconfiguration cannot keep up with the dynamically changing personalized market demand,a large number of high-quality manufacturing resources after the physical production line reconfiguration are idle and shut down,resulting in a huge waste of manufacturing resources.Therefore,how to effectively integrate idle manufacturing resources to dynamically and rapidly reconfigure personalized customized virtual production lines(calling virtual production lines in this paper)has become an urgent problem for the manufacturing industry.In response to the above problems,based on investigating domestic and foreign theories on reconfigurable manufacturing systems,the research on virtual production line reconfiguration based on idle resources of multi-physical production lines includes:(1)Aiming at the shortage of multi-physics production line’s ability to carry out refined collaborative production,a new cloud-edge network collaborative control mode based on hierarchical multi-granularity digital twins—Fog-Manufacturing(FMfg)was proposed.Self-similarity Manufacturing Components(SMC)conduct self-organization,layering,and on-demand collaboration to realize rapid reconfiguration of the production system for personalized orders in a multi-physical production line environment.In FMfg environment,in view of the difficulty of decoupling physical production lines,a fractal modeling method for manufacturing components in FMfg and its evolution mechanism based on a multi-agent deep reinforcement learning model were proposed.In order to meet the production needs of dynamic,personalized production,an SMC-based virtual production line reconfiguration and optimization architecture in FMfg was proposed.(2)Aiming at the optimization problem of virtual production line reconfiguration for a new kind of workpiece set,the matching degree between the production process of that workpiece set and each single physical production line was analyzed,and the physical production line with the highest matching degree was selected as the basic production line.Search the processes that cannot be processed for their available manufacturing resources in other physical production lines.Take the manufacturing cost and the product quality as optimization goals;consider various mandatory time constraints such as delivery time,equipment occupancy,etc.,and establish a virtual production line reconfiguration optimization model with considering time constraints in a multi-physical production line environment.To solve that model,an improved Harris Hawks Optimizer(HHO)was proposed.Experimental verification was carried out,realizing the multi-objective optimal configuration of virtual production line reconfiguration for a single workpiece in FMfg.(3)In the multi-physical production line environment,aiming at the optimization problem of virtual production line reconfiguration of multiple types of workpiece set,the competition process of multiple types of workpiece set for the critical resource and the actual constraints involved were analyzed.As a result,completion time,manufacturing cost,product quality were regarded as the multi-optimization objectives,a multi-objective optimization model for virtual production line reconfiguration considering the competition of the critical resources was constructed.To solve that model,an improved whale optimization algorithm was proposed.Experimental verification was carried out,which realizes the rational and effective utilization of critical resources,and obtains virtual production line reconfiguration optimization methods that are suitable for multiple workpiece set orders.(4)Aiming at the reconfiguration optimization problem of disturbed virtual production lines,the actual constraints in the reconfiguration process of concerned production lines were analyzed.The manufacturing time,manufacturing cost and the comprehensive product quality are the multiple optimization objectives,and a multi-objective optimization model of disturbance production line inheritance reconfiguration was established with considering production disturbances such as equipment failures.A multi-objective whale adaptive optimization algorithm with multi-swarm adaptive parameter adjustment capability was proposed.Experimental verification was carried out,where inheritance adaptive reconfiguration optimization configuration of virtual production line under internal and external disturbance environment was realized.(5)In order to integrate and verify the theoretical models and algorithms proposed in this paper,a personalized customized virtual production line reconfiguration system in FMfg was developed.That system with the theoretical model and algorithm proposed in this paper has been initially applied and verified in "Research on the theory and method of Complex Manufacturing Cooperative Control and Decision Making based on Humancyber-machine fusion".Good results have been achieved.