3e2a1b56-dafb-454d-87ad-86adea3e7b86

290
Part Three
Planning and control
Figure 10.15 Push versus pull: the gravity analogy
own suppliers. In this way, demand is transmitted back through the stages from the original
point of demand by the original customer.
The inventory consequences of push and pull
Understanding the differing principles of push and pull is important because they have
different effects in terms of their propensities to accumulate inventory in the operation. Pull
systems are far less likely to result in inventory build-up and are therefore favoured by JIT
operations (see Chapter 15). To understand why this is so, consider an analogy: the ‘gravity’
analogy is illustrated in Figure 10.15. Here a push system is represented by an operation, each
stage of which is on a lower level than the previous stage. When parts are processed by each
stage, it pushes them down the slope to the next stage. Any delay or problem at that stage
will result in the parts accumulating as inventory. In the pull system, parts cannot naturally
flow uphill, so they can only progress if the next stage along deliberately pulls them forward.
Under these circumstances, inventory cannot accumulate as easily.
Drum, buffer, rope
Drum, buffer, rope
Theory of constraints
The drum, buffer, rope concept comes from the theory of constraints (TOC) and a concept
called optimized production technology (OPT) originally described by Eli Goldratt in his
novel The Goal. 8 (We will deal more with his ideas in Chapter 15.) It is an idea that helps to
decide exactly where in a process control should occur. Most do not have the same amount
of work loaded onto each separate work centre (that is, they are not perfectly balanced. This
means there is likely to be a part of the process which is acting as a bottleneck on the work
flowing through the process. Goldratt argued that the bottleneck in the process should be
the control point of the whole process. It is called the drum because it sets the ‘beat’ for the
rest of the process to follow. Because it does not have sufficient capacity, a bottleneck is
(or should be) working all the time. Therefore, it is sensible to keep a buffer of inventory
in front of it to make sure that it always has something to work on. Because it constrains
the output of the whole process, any time lost at the bottleneck will affect the output from
the whole process. So it is not worthwhile for the parts of the process before the bottleneck
to work to their full capacity. All they would do is produce work which would accumulate
further along in the process up to the point where the bottleneck is constraining the flow.