Extending the Scope of Analytical Methods for Performance Evaluation of Manufacturing Flow Systems

Manufacturing flow systems consisting of machines separated by buffers of finite capacity have been studied a lot in the literature because systems of this type are widely encountered in industry. The basic assumptions are that the processing time of each machine is deterministic and machines are subject to random failures. Failure and repair times of each machine are characterized by probabilistic distributions. The corresponding model is referred to as the asynchronous model. The basic configuration that has been widely studied is the so-called production lines. Analytical methods for performance evaluation of production lines have been developed for two different models: the synchronous model and the continuous flow model. For the range of parameters usually encountered in practical applications, the behavior of either model is close to that of the asynchronous model, and therefore, these models can be used as approximations of asynchronous models. Moreover, for both models, efficient analytical approximation technique have been developed. Most of them methods are based on the decomposition method originally proposed by Gershwin and further developed and improved by him and other researchers. Such decomposition methods have been extended to other topologies, namely closed-loop production lines and assembly/disassembly systems. However, there are other topologies and/or features that are encountered in real systems for which very little work on analytical methods has been done so far. In this paper, we show that some of these features can easily be incorporated into analytical methods based on the continuous flow model. These features include random yield, split and merge mechanisms, and multi-items assembly systems.