Abstract: To draft dynamic scheduling strategy based on Networked Manufacturing (NM) among supply chain cooperators,a three-layer supply chain dynamic scheduling model composed of manufacturer,cooperative manufacturer and distributor was constructed.Aiming at the multi-product and multi-order requirement,the scheduling model consisted of production costs,semi-production inventory costs,product inventory costs,transportation costs and delay punishment costs was built to search the optimal operation time,costs and customer service level under the constraints of delivery date and process sequence.To solve the problem that the long solution time was caused by full permutation of process sequence,the single species compound ant colony algorithm was used to judge order types with main console,and the adjacent exchange factor was put forward to adjust order sequence for improving the optimization efficiency.According to the parity of adjacent exchange frequency,the forward strategy of orders was determined to avoid dead loop.The combined transportation strategy was considered for the same distributor,and the transportation combination could be achieved as a special process.The simulation results showed that the proposed algorithm had good solving performance and search speed,and the constructed scheduling model was reasonable and practical,which could provide reference for real production scheduling.

Key words: multi-product supply chain scheduling, networked manufacturing, adjacent exchange algorithm, compound ant colony algorithm, combined transportation

摘要： 为制定网络化制造模式下供应链合作成员间的动态调度策略,构建了由制造商、协同制造商和分销商组成的三层供应链动态调度模型。针对多产品多订单调度情况,在加工工序和交货期等约束条件下,构建由加工成本、半成品库存成本、成品库存成本、运输成本和延期惩罚成本等描述的调度模型,以寻求最优化的供应链运作时间、综合成本和客户需求服务水平。为解决多产品多订单加工顺序全排列与求解穷搜索所带来的时间过长的问题,采用单种群复合蚁群算法,运用总控制台判断订单类别以分配至待加工订单库,提出相邻交换因子调整订单顺序以提高蚁群寻优效率,并根据相邻交换的奇偶次数决定订单前移策略,避免出现死循环|同时,在供应链调度过程中考虑相同分销商的不同产品订单是否进行组合运输,通过将订单合并行为作为一特殊工序加以实现。实例仿真表明,该算法具有良好的求解性能和搜索速度,且调度模型具有合理性和实用性,可以为实际生产调度的优化决策提供参考。

关键词: 多产品供应链调度, 网络化制造, 相邻交换算法, 复合蚁群算法, 组合运输