One-for-one period policy and its optimal solution over a finite horizon

Recently, a new ordering policy named one-for-one period policy has been introduced in a steady state condition for the zero ordering cost with an assumption of lost sales. In this policy, constant time interval between two consecutive unique orders is assumed. In contrast to this policy, this paper addresses a new approach in which inter-arrival times are determined in a finite horizon with limited amount of arrivals. Due to the transient condition of our approach, namely (S(n), 1), matrix multiplication must be employed, but it quickly becomes cumbersome for large n as there are (n − 1) decision variables for n arrivals. Thus, we have invoked the genetic search strategy to reduce the amount of search. Finally, we provide a numerical analysis to evaluate the performance of our approach. The results showed that by applying the suggested approach we can save cost compared with the one-for-one period policy for n < 200, especially when the ratio of lost sales to holding cost is large. Furthermore, arrivals scheduling creates a dome-shaped inter-arrival times.