Research On Dynamic Reconfiguration And Scheduling Problems Of Virtual Manufacturing Cell Towards Random Orders

Virtual cellular manufacturing system is an advanced organization mode of production which is gradually developed from traditional cellular manufacturing on the basis of group technology. Based on the dynamic reconfigurable characteristics, the manufacturing system is capable of responding to changes in product demand quickly, and adapt to the current production mode of many varieties with small quantities flexibly. The main advantage is that when the type or quantity of production task is changed, it is unnecessary to change the physical location of equipment resources, what needs to be done is cell reconstruction only in a logical way. This method can respond to changes and is helpful for manufacturers to reduce production adjustment costs and time, improving production performance greatly.With the production rhythm to speed up, the current market environment showing a feature of dynamic and uncertainty, but research on dynamic virtual cellular problem is still not perfect, from the perspective of reconfigurable manufacturing systems, and make-to-order as a starting point, this article carried out a detailed study on dynamic reconfiguration and scheduling problems of virtual cellular with orders arriving randomly.In this paper, research on this issue is divided into two stages. The first phase of the study is the real-time reconfiguration of virtual cells under conditions of new order randomly introduced. First, based on the similarity degree between new tasks and existing virtual cell, neural network algorithm is carried out to pre-allocate all tasks in new order, it is prior to consider assigning the new task into an existing production cell, to maximize continue the existing cell structure, thereby reducing disturbance to existing production scheduling plan, otherwise reconstruct a new cell. After the orders are assigned into cell manufacturing system, then optimize the system according to the target of minimum number of cross- cell processing, cell load balancing and cell adjustment range, using heuristic algorithm to optimize the reconstruction program, in response to changes in real-time while maintaining stable operation of the system. The first phase complete the development of the virtual cell reconfiguration program from the level of organization of resources, and the second stage is from the level of manufacturing execution to further study on virtual cell rescheduling problem after new tasks added. In the framework of rolling horizon decomposition method, an improved event-driven and periodic mixing rescheduling mechanism is proposed. By periodic driven mechanism to tracking the change of system, and the task urgency judgment strategy which is based on the rolling cycle anddelivery date is added into the event-driven mechanism, as to avoid unnecessary rescheduling, taking into account the system’s flexibility and stability at the same time.Virtual cell rescheduling mathematical model among rolling window was established to minimize the target of total completion time and transport distance, using discrete particle swarm algorithm mixed genetic manipulation to solve the model and achieved satisfactory results.In order to verify the application of the proposed method in practical enterprise, the paper has carried in-depth research on a ship Manufacturing enterprise, detailed analysis of current situation and problems of the enterprise’s production management has been made,and apply this design scheme to the shipyard outfitting pipe production process, we proposed improved program about reconfiguration scheduling problem of virtual cell with new order added, by comparing the results of designing programs before and after improvement, we can confirm the feasibility and effectiveness of the methods mentioned in this article.