Schedule-based material requirements planning (SBMRP), which has been suggested as an integration of scheduling and MRP methods, is described and its deficiencies are identified. Knowledge acquisition through computer experimentation (KACE) was used to develop a knowledge-based expert system (KBES) for SBMRP. Comparison of the KBES with conventional forward- and backward-method SBMRP, using total cost as a criterion showed the KBES to be superior for scheduling jobs consisting of up to 20 parts. Sensitivity analysis was conducted on the important parameters such as total available time over total processing time (TA/TP), processing time for each operation, and number of parts. The KBES exhibited the best performance regardless of the processing time. As the level of TA/TP increased, the KBES's performance improved.
展开▼