可重入作业车间排队网建模与缓冲区配置优化

Queuing network modelling and optimal buffer allocation for reentrant job-shops

针对定制型制造企业可重入作业车间,研究了排队网建模和缓冲区容量配置优化方法。首先,建立了具有任意拓扑结构的有限缓冲区开排队网模型,以分析作业车间生产活动的随机过程,将节点的输入/输出过程拟合为一般分布,并提出近似求解的速率迭代法,用以快速计算系统平均产出率和生产周期。其次,通过算例实验与仿真结果进行对比,验证了假设模型的有效性和求解方法的精确性,并分析了缓冲区容量对系统性能的影响。缓冲区容量配置优化问题是一个非线性整数规划问题,也是一个NP难组合优化问题,因此提出一种由速率迭代法与Polyblock算法嵌套迭代的优化方法,并通过案例结果进行了验证。

In the reentrant job-shop of customized manufacturing industry,queuing network modelling and buffer allocation problem were studied.The system was modelled as an open queuing network with arbitrary topology to describe the random process of production activities.The input and output processes were assumed to be general distributions.An approximation method called rate iterative method was proposed to compute the system mean throughput and cycle time.The efficiency of the hypothetical model and the accuracy of the solution method were tested by comparing with simulations from some experiments.The optimal buffer allocation problem was a nonlinear integer programming problem and an NP-hard combinatorial optimization problem,Anested iteration optimization algorithm by rate iterative method and polyblock method was developed to solve it,and the cases were used to verified.