Asymptotic analysis and synthesis of serial production systems.
F. Top · Deep Blue (University of Michigan) · 1990
Asymptotic models of synchronous and asynchronous serial production systems are introduced and analyzed. The results obtained indicate that these asymptotic models are, on one hand, simpler for analysis than traditional models discussed in the literature for the last 25 years and, on the other hand, are sufficiently rich to reflect the behavior of real manufacturing systems. As far as the analysis is concerned, an asymptotic theory is developed that leads to the calculation of the average production rate and buffer occupancies as functions of systems parameters. Therefore, the theory presented here can be used an an inexpensive tool for analyzing manufacturing processes in mass production systems eliminating, in some cases, the necessity of complex and costly computer simulations. From the point of view of the synthesis of synchronous serial production systems, the dissertation considers three design problems: (1) the optimal workforce assignment problem, (2) the simultaneous workforce and buffer capacity assignment problem and (3) buffer capacity assignment problem. It is shown that the optimality in design problems (1) and (3) is achieved if and only if the so-called dynamic buffer and machine balancing conditions, derived in the thesis, are satisfied, respectively. The optimality in design problem (2), however, is ensured by either of the dynamic balancing conditions and the optimally designed (dynamically balanced) system satisfies both conditions. It is generally recognized that the synchronous serial production lines are advantageous from the production rate point of view whereas the asynchronous ones ensure a higher quality. Both, however, are usually run in a balanced mode where the cycle times of all machines are identical. In the present dissertation, it is shown that the asynchronous lines, being run in an appropriate unbalanced mode, could have a production rate close to that of the synchronous lines. To this end, the dissertation formulates the problem of optimal cycle time assignment and gives a solution to this problem. It turns out that, in this problem too, the optimality is achieved if and only if the dynamic buffer balancing condition is satisfied. The findings of the thesis are then applied to a case study of a paint shop system in a modern automobile assembly plant, and the improvement in the average production rate, predicted theoretically and observed practically, is reported.