基于改进降顶算法的非可靠生产线最优缓存分配

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.

Improved degraded ceiling algorithm for optimal buffer allocation in unreliable production lines

Buffer allocation is a major optimization problem in manufacturing systems design. The recursive decomposition model and its statistic characteristics are studied in the unreliable successive product firstly, and the Monte-Carlo simulation method is used to estimate the throughput of product line in the condition of fixed buffer allocation. Comparing with the conventional estimation method, it is easy to depict the production conditions. Then, the tabu set based improved degraded ceiling algorithm is proposed to solve the optional buffer allocation problem. Because of the tabu set with memorability, the search efficient of the algorithm can be increased greatly. Finally, the computer simulation results show that the proposed algorithm can search for the optimum solution effectively and fleetly.