TOC has a Five Step Process of On-Going Improvement:
- Identify the system constraint.
- Decide how to exploit the constraint.
- Subordinate everything else to the above decision.
- Elevate the constraint.
- Avoid inertia. Go back to Step 1 and determine if the constraint has moved.
There is a Step 0 that is almost always done first. Choke the release of orders to the shop to reduce Work-In-Progress (WIP). This by itself will reduce lead-time dramatically.
Exploiting the constraint looks for ways to improve its performance, such as reducing downtime for lunch and breaks; giving it priority for repair; not having it work on non-priority or defective raw material.
Subordinating requires the rest of the system to be synchronized to the pace of the constraint. This is a paradigm shift for most people. For example, it's okay to be idle if there is no work to be done. Elevating means obtaining more constraint capacity. This is the first point at which money is spent on the system.
Avoid inertia to keep the system improving. Go back to Step 1 to determine if there is a new constraint.
Typical manufacturing lead times have only 5 to 10% processing time. The rest of the time, orders are waiting in queues. The constraint of a manufacturing system is called the Drum resource. It is the most heavily loaded resource. Therefore, it should be the drumbeat to set the pace of the system.
A mechanism to release new work into the system is called the Rope. As the Drum completes work, the Rope allows new orders to be released. In order for the Drum to always have work, a Buffer of work is maintained in front of the Drum to protect it from starving. If the Drum stops working, the system throughput is stopped. An hour of downtime for the Drum is an hour of lost throughput to the system. Watch this drum resource carefully. Protect it.
If the Drum (constraint) capacity is increased, the system capacity is immediately increased. If the Drum capacity is raised above the next most loaded resource, that resource is now the Drum.
As a very simplistic example, the system shown in Figure 2 has five resources in sequence with capacities of A=200/hour, B=180/hour, C=150/hour, D=170/hour and E=190/hour. It could be a production line or any process. What is the maximum amount this system can yield in an hour?
Figure 2: Production sequence with different production rates
Increasing the capacity of B to 200/hour will not increase the system output.
C is the constraint limiting the system. It's used as the drumbeat of the system. All the work-in-progress (WIP) after Raw Material and before the Constraint is considered Buffer. The Rope, see Figure 3, is a device, a method, to monitor the Buffer and advise Raw Material to release more work to refill the Buffer.
If A produces 200 per hour, work will pile up in front of B, which can only process 180/hour, since C can only process 150/hour. Both A and B must be subordinated to the constraint C. They must keep C supplied with work; yet not bury it with work. They maintain C's Buffer in a predetermined range. If A or B breaks down, they have the extra capacity to refill the Buffer. If C, the constraint, breaks down, A and B sit idle until the constraint is back on line.
Figure 3: The sequence showing Drum, Rope and Buffer
Figure 3 shows that the drumbeat is 150/hour as emphasized in Figure 4. This is the rate at which we can make money. Every machine can work at its capacity when it has work. When it does not have work, it can't work. If B stops working, C will continue working, using the Buffer. The Buffer is sized to accommodate reasonable variation in the system. When B starts working again, since it is faster than C, it will refill the Buffer.
Figure 4: The production chain showing the system's constraint