The impact of uncertain yield on capacity acquisition in process plant networks

In this paper, we consider the case of process plant networks, which consolidate production of components. It is well known that such networks take advantage of process commonalities thereby accruing economies of scale advantages. It has also been shown in recent work that besides scale economies, process plant networks also enjoy risk-pooling advantages which are a direct outcome of network structure. In this research, motivated by real-world examples, we study capacity acquisition in a process plant network when demand and process yield are both uncertain. We show that the optimal capacity acquired in such situations may be considerably higher or lower than that acquired when yield is assumed perfect. We develop distribution-free policies for optimal capacity acquisition. We illustrate our results by means of a numerical example and study the sensitivity to important model parameters. For specific distributional assumptions we also provide closed-form results for the optimal capacity acquisition. We consider our work as perhaps the first to bridge the gap between operational level quality control studies and strategic level network configuration research. Overall, we believe that this paper opens up several avenues for future research integrating these two important areas.