Echelon-stock (R,nTR,nT) control in two-stage serial stochastic inventory systems

For the two-stage (R,nTR,nT) policy, we present an algorithm to estimate expected costs, given (R,nTR,nT), and to compute near-optimal (R,nTR,nT) values. We illustrate that the (R,nTR,nT) policy is cost-effective compared to an (r,nQ)(r,nQ) policy and to lower bounds on minimum achievable costs. We explore the impact of parameters on costs.