A stochastic assembly system with resume levels

In small-lot, multi-product and multi-level assembly systems, kitting (accumulating components required for assembly) plays a crucial role in determining system performance, especially when the system operates under a stochastic environment. This paper analyzes the kitting process of a stochastic assembly system, treating it as an assembly-like queue. Two types of components which are required to complete a kit, independently arrive at the buffers according to Poisson processes. Furthermore, the arrival flow is shut down when the queue size attains the buffer size until it decreases to a specified level, namely resume level. Using a Markov renewal approach, we derive the exact distribution of the kit completion time interval. The distribution of the sojourn time in the buffer of each component and the loss probability of components are also obtained. Finally, we show the merit of resume levels via numerical examples.