Abstract The design of a production line directly affects the system performance which is usually measured by its throughput. The problem involving determination of the optimal capacity and location of the buffers in a production line is known as the buffer allocation problem (BAP). Due to the difficulties such as the NP-hard structure of the problem and not being able to be defined the throughput of the line in terms of the buffer capacities algebraically, meta-heuristic search algorithms are widely used to solve the BAP. In this study, an adaptive large neighborhood search (ALNS) algorithm is proposed to solve the BAP for throughput maximization in unreliable production lines. Different from the literature, for the first time, ALNS algorithm is employed to solve the problem of designing a production line. For this purpose, two different removal-insertion operator pairs are proposed and employed in an adaptive way by considering the nature of the problem. Moreover, a new initialization procedure based on the well-known storage bowl phenomenon concept is proposed to reduce the search effort. Performance of the proposed algorithm was tested on the existing benchmark instances. A computational study demonstrated the benefits of not only the adaptive mechanism embedded into the proposed algorithm but also the proposed initialization procedure.

中文翻译：

---

不可靠生产线中的缓冲容量分配：一种自适应大邻域搜索方法

摘要 生产线的设计直接影响系统性能，系统性能通常以生产量来衡量。涉及确定生产线中缓冲区的最佳容量和位置的问题称为缓冲区分配问题 (BAP)。由于问题的 NP-hard 结构以及无法用代数方式定义线路的吞吐量等困难，元启发式搜索算法被广泛用于解决 BAP。在这项研究中，提出了一种自适应大邻域搜索 (ALNS) 算法来解决不可靠生产线中吞吐量最大化的 BAP。与文献不同，首次采用ALNS算法解决生产线设计问题。以此目的，考虑到问题的性质，提出并以自适应方式采用了两种不同的移除-插入算子对。此外，提出了一种基于众所周知的存储碗现象概念的新初始化程序，以减少搜索工作量。在现有的基准实例上测试了所提出算法的性能。一项计算研究证明了不仅嵌入到所提出的算法中的自适应机制而且所提出的初始化程序的好处。