Flexible assembly and shortest queue problems

For the design of flexible assembly systems an iterative approach seems advisable. Such an approach would require the evaluation of several alternative designs. Therefore it is necessary that individual designs can be evaluated very efficiently. In this paper it is argued that extended shortest queue models can play an important role in designing flexible assembly systems. Consequently, it would be essential to have efficient methods to evaluate shortest queue models. The paper presents a new approach for shortest queue models and shows how this approach can be used for several variants. The approach uses mixtures of product forms as approximative solutions for the equilibrium distribution of the Markov chain. Although no rigorous proofs are presented in this paper, it is indicated that the exact solution can be approximated deliberately good in this way with the particular property that the approach also generates upper and lower bounds.