In this paper, we consider a discrete flow model with transportation delays for more realistic performance evaluation and optimization of failure-prone manufacturing systems. This model is applied to a manufacturing system composed by a single machine, a buffer and a stochastic demand. The goal of this paper is to evaluate the good buffer level taken into account the transportation time, the machine failures, the inventory cost, the transportation cost and the lost sales cost. The estimators of the difference of the total expected cost are proven to be unbiased and a simulation algorithm is proposed to evaluate the performance of the manufacturing system in the presence of transportation delay.