论文部分内容阅读
缓存分配是生产系统设计中一个重要的优化问题.基于非可靠连续生产线递推分解方法及其统计特性,提出在给定缓冲配制条件下的蒙特卡洛生产线吞吐量仿真估算方法,相比于传统的吞吐量估算模型该方法能更准确地描述各种生产场景.通过构造具有记忆性的禁忌集改进了传统降顶算法,并将其应用于实际缓存最优分配方案搜索中.仿真结果表明,对于各种规模的平衡生产线和非平衡生产线,改进降顶算法都可以快速有效地搜索到最优解.
Caching allocation is an important optimization problem in production system design.Based on recursive decomposition method and its statistical characteristics of non-reliable continuous production line, a method to estimate the throughput of Monte Carlo production line under a given buffer formulation is proposed. Compared with traditional This method can describe all kinds of production scenarios more accurately.It improves the traditional descent algorithm by constructing a taboo set with memory and applies it in the search of the optimal cache allocation scheme.The simulation results show that, For all sizes of balanced and unbalanced production lines, the improved descent algorithm can quickly and efficiently search the optimal solution.