It is difficult to control spare parts inventory of a product to meet certain maintenance demand. The problem becomes more challenging when the installed base of the product changes over time. Under this situation, the inventory value needs to be adjusted according to the resulting non-stationary maintenance demand. This challenge is usually encountered when a manufacturer starts selling a new product and agrees to provide spare parts for maintenance. In this paper, a special case involving a new non-repairable product with a single spare pool is considered. It is assumed that the rate of new sales of the product is constant, and the product's failure time follows the Weibull distribution. The mathematical model for the resulting maintenance demand is formulated and calculated through simulation. Based on the maintenance demand, a dynamic (Q, r) - (lotsize/reorder-point) restocking policy is formulated and solved using a multi-resolution approach. Finally, a numerical example with the objective of minimizing the inventory cost under a service level constraint is provided to demonstrate the proposed methodology in practical use.