he problem of order-release planning for discrete parts manufacturing is addressed. Traditionally, order-release planning in a multi-stage shop is performed with material requirements planning (MRP) logic, which generally assumes that capacity is infinite and that lead times can be estimated using historical data, past experience, and general rules-of-thumb. These assumptions often result in infeasiblc plans. A new approach to order-release planning termed BACKSIM, that is based on discrete simulation, is proposed. With this approach, order-release plans are generated via a backward explosion logic similar to MRP except that a simulation model of the facility is used. Since the simulation model accurately captures real-world capacity constraints and operational rules, the order-release plans are theoretically feasible. Essentially, component lead times are calculated based on the anticipated queueing in the system each time BACKSIM is executed. A technique is developed that allows BACKSIM to be fairly compared with the MRP approach.