面向订单基于Agent的启发式产品组合生产计划研究

由于需求的不确定,很多企业在综合生产决策中渐渐采取面向定单的生产计划方式.当订单需求大于企业最大生产能力时,需要进行产品组合决策,以确定使利润最大的多产品混合生产的产量.本文针对一个面向订单生产的半离散制造型企业,结合Agent,建立了一个不依赖于特定成本和利润函数形式的基于作业流程和BOM结构的分布式生产决策模型,并给出了基于粒子群优化算法的启发式求解方法.该算法通过一维搜索调整粒子群"飞翔"的速度系数以解决寻优过程中粒子"飞出"可行域的问题,通过算例验证了求解算法的有效性.为将算法融合到所研究企业的ERP系统中,开发了试验性原型系统,并进行了模拟决策,结果表明所提出的模型和算法是可行的.

A Research on Agent-Based Heuristic Production Planning of Product Mix Considering Build-To-Order

For the uncertainty of demand,many enterprises have adopted the production planning of Build-To-Order(BTO).If demand in the order exceeds the enterprise's production capacity,the product mix decision should be made for achieving the optimized production level that makes the maximized profit.In a semi-discrete manufacturer that adopts BTO,integrating Agent,the paper built a distributed production decision model based on activity processes and the Bill of Materials(BOM).The model is independent of the specific formation of cost and profit function.A heuristic algorithm has been proposed to solve the model,which is based on Particle Swarm Optimizer(PSO).And the algorithm uses one dimension search to adjust the flying velocity coefficients of particles to control particles not to fly out of the search space.Numeric examples research demonstrates that the algorithm is effective.For applying the model in a real ERP system,a testing prototype has been developed and a simulating decision was made.The results show that the decision model and the algorithm are feasible.