Research Of Multi-period Virtual Cellular Formation And Dynamic Scheduling Under Sequential Orders Arriving

Manufacturing industry gradually adapt the Make-to-Order production mode in an increasingly competitive and manufacturing capability improving environment, in which the product present the feature of small batch, species diversity and dynamic variable market demand, etc. Virtual cellular manufacturing system is a widely used production organization way in enterprise with many variety and small batch product, which combines the efficiency of cellular manufacturing and the flexibility of flexible manufacturing. VCMS can not only guarantee the processing efficiency, but also greatly reduced the time and cost of reconfiguration through the logical cell formation organization production, therefore more available to the enterprise facing the dynamic variable internal and external production environment.Except the outside disturbance such as emergency order, the planning level will release ordinary order constantly so that the orders will be processed and completed in turn in the actual production environment. Because of the sequential arrived production task and complicate variable order group, lead to the cell structure formatted with the part type and demand of the last stage may not be fit for the next stage. Under virtual cellular production mode, the original plan horizon should be decomposed into multiple consecutive production periods, and the internal resources in production system should be adjusted constantly according to the dynamic changes. However, frequently reconfiguration will lead to more material handling costs and make the production scheduling be more complex, especially greatly increases the pressure of process operator, which will bring many hidden dangers to the production process control and quality control. So it is necessary to explore a dynamic virtual cellular formation method considering the multi-period equilibrium strategies and subsequent scheduling complexity. It will reflect the actual production situation well to consider the production order is arrived sequentially in a certain production cycle, but it will make the scheduling process be more randomness and increase the complexity of the scheduling problem. So the periodic driven dynamic adjustment strategy combined the technology of rolling window should be adapted to simplify the dynamic scheduling, making the sequential arrived order into processing in the rational sequence, and the equipment utilization will be improved by the resource share between virtual cell.This paper is from the perspective of the short-term product plan in workshop level, the virtual cellular multi-period formation problem is divided into two stages of multi-period resource selection and distribution of process route with cell structure optimization. The processing batch with equipment redundancy, system robustness and the compact of cellular formation to scheduling must be considered comprehensively to select the manufacture resource according to the different of part type and batch in different production cycle. A multi-object dynamic mixed integer programming model is designed to complete the source selection and route batch assignment by minimum the process and transfer cost, schedule factor and load balance. The optimal pareto solution will be achieved by adaptive genetic algorithm and the best source group will be determined based in TOPSIS. Then an improved source cluster algorithm based in the fuzzy C-means is presented to complete the part and equipment clustering, which introduce the HERBAL algorithm to generate the initial cluster center and adapt the specified distance function of cellular formation to cluster the source after selected. The clustering program will be evaluated from group efficiency and load balance between equipments to achieve the optimal virtual cellular multi-period formation.To the sequentially arrived orders in a cycle, through the simulation on random arrived time of orders and the combination of resource share between virtual cells, a dynamic scheduling strategy considering the equipment real-time load based in variable period is proposed in this paper, which figures each processing tasks sequentially arrived into the corresponding scheduling interval. A multi-objective optimization model is constructed to this scheduling problem, in which the max lead time, weighted total tardiness and average process time of each part are considered. The improved ant colony algorithm is presented to solve the problem in using the maximum minimum ant mechanism to avoid local optimum, and the elite ant strategy is introduced to accelerate the convergence of the algorithm, in which the dynamic virtual cellular scheduling solution is achieved.In order to verify the effectiveness of method presented in this paper, which include virtual cellular resource selection, clustering and dynamic scheduling in multi-period, after the detailed analysis the production status and existed problem in a shipbuilding corporation, applying the method presented to a pipe processing workshop. Through the comparison between current production mode and virtual cellular manufacturing mode, in which shows the feasibility and effectiveness of the proposed method.