Multi-type products and dedicated buffers-based flexible production process analysis of serial Bernoulli lines

In the contemporary manufacturing environment, flexible production is growing in popularity due to its quick adaptation and response to the market changes and customized requirements. Since multi-type products under small-batch-based operations are commonly seen in flexible production, dedicated buffers are sometimes used to separate different types of parts that are being processed on one production line. Besides, the steady-state analysis may be ineffective in analyzing such systems because of the small-batch-based operation. As a result, transient-based production process analysis becomes indispensable in studying the dynamic behavior of flexible production systems. In this work, we investigate the production process in flexible production systems with multi-type products, machines obeying the Bernoulli reliability model, and dedicated buffers. In the first place, based on Markovian analysis, a set of auxiliary production lines are presented to decouple the dynamics of the original system. Next, the process performance measures of interest are approximately evaluated from the auxiliary production lines to avoid the issue of system state explosion, which is frequently encountered when applying the Markovian approach. Besides, the proposed analytical algorithm is validated to be with high precision through numerical experiments in contrast with the Monte Carlo simulation. Finally, by using the proposed method and applying the genetic algorithm, we investigate the problem of buffer allocation to minimize the completion time of the system.