MRP System Built for Real Manufacturing

Why an MRP Plan on Its Own Is Not a Plan You Can Build

Classic MRP assumes infinite capacity. It takes a due date, subtracts a fixed lead time per level of the bill of material, and tells you when to release. It never asks whether the work centers have the hours. So the purchase orders are right, the dates are wrong, and the shortage list and the expediting start in week three.

A worked example. Say you need 200 assemblies on day 20. The bill of material has two machined parts per assembly, so 400 parts, and the mill runs them at 6 minutes each: 2,400 minutes, 40 hours of mill time. MRP sees a 5-day purchase lead time for the bar stock and a 3-day standard lead time on the machining step, so it releases the order on day 12 and reports the plan as feasible.

Now look at the mill. It is staffed one shift, 8 hours a day, and it already has 26 hours of other work booked in that window. The job has to be finished by day 20, so the eight working days from day 12 to day 19 are what it gets: 64 hours of capacity, 38 of them free. The job needs 40. It does not fit, by two hours, and nothing in the material plan knows that. Finite capacity scheduling catches it while there is still time to move a job, add an hour of overtime, or tell the customer before the date is missed.

That is the whole difference between MRP as arithmetic and MRP as a plan. In EDGEBIC Complete the material requirements and the capacity schedule are the same model, running on the RMDB finite capacity engine that has been scheduling plants since 1991, so the dates MRP writes back are dates the floor can keep. The method behind it is finite capacity scheduling, and the gross-to-net arithmetic itself is worked through in the MRP net requirements formula.