Advanced planning systems and MRP based planning part 2

VW 2021-03-16

To continue on my previous post: several things were happening in the 1970s-1990s, when the term MRP was replaced by MRP-II and later by ERP. For clarity, let's summarize the terms:

MRP (I): Material Requirements Planning
MRP II: Manufacturing Resources Planning
ERP: Enterprise Resources Planning
APS: Advanced Planning and Scheduling (1

1) note that Gartner and other parties have replaced the term APS by SCP&O, which stands for Supply Chain Planning & Optimization, but I prefer to stick to the old term APS.

MRP II basically is MRP I plus a number of extensions, to compensate for the fact that MRP does not take capacity constraints into account very well. Below is a picture of the MRP II framework, with MRP (I) in the middle. This framework still is the basis of the planning functionality of ERP systems today.

MRP II contains tools such as:

Rough-cut capacity planning (RCCP), which provides a high level capacity check on the Master Production Schedule level.
Capacity resource planning (CRP), which introduces a basic way to visualize potential capacity problems resulting from the MRP-I run.

Though RCCP and CRP were introduced to deal with the shortcomings of MRP, they did not much more than visualize potential capacity issues. Solving such issues was still the task of the human planner. For example, when CRP would show an overload on a specific workcenter, the planner would have to manually replan order by order to solve the overloads. Initially, no Gantt charts nor interactive tools were offered.

In parallel of these developments, the first so-called 'finite capacity planning systems' were introduced, containing an electronic Gantt chart, as computers enabled system designers to create graphical user interfaces. In Germany, such systems were known as Leitstands (control posts) and basically were an add-on to databases containing orders or jobs. Commercial systems for planning and scheduling that can be seen as the predecessors of current APS systems were introduced in the 1980s, initially known as finite capacity planning (FCP) systems. The name FCP indicates that these systems, unlike their MRP counterparts, enabled taking finite capacity into account. Such systems were typically implemented as an add-on to MRP systems, importing the results of the MRP run to create a schedule.

MRP, ERP, FCP and APS are all produced by practitioners, trying to solve real problems. But where were the innovations from science? Researchers were dealing with planning and scheduling problems that were incredibly simple, to be able to apply math to it. Thousands of papers were written on problems with for example two machines, a few jobs, known arrival times - assumptions that come nowhere near a real-life problem.

However, since about the 1980s, an increasing number of researchers realized that to tackle any realistic problem, an analytical approach would have to be dropped in favor of techniques that search for a “good” (but not optimal) solution. Computer power was getting cheaper, and thereby such search techniques were becoming more and more a feasible approach. Effort was invested in designing search techniques that would find a good solution within reasonable time, as the solution space for more realistic problems is very large. The gap between theory and practice was getting somewhat smaller as a result, as researchers used problems from practice to create their models.

To be continued in a future post.