Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the

copyright owner

Chapter 15 - THEORY OF CONSTRAINTS IN

OPERATIONS
LEARNING OBJECTIVES
● Know the concepts of the Theory of Constraints and their possible impact on the
management of capacity and performance of operations.
● Understand how the Theory of Constraints changes the traditional approach to
analyze capacity and flows in operations and what the implications of this
change are for operations management.
● Understand the principles of Theory of Constraints and how they can be used in
the management of physical goods and service operations.
● Understand the workings and decision-making process of the Theory of
Constraints in relation to scheduling and controlling operations.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

15.1 INTRODUCTION
This chapter will change the way you view operations, capacity management of
operations systems, and make you a much more critical observer of queue systems and
flows of people, materials, and information. The concepts behind the "Theory of
Constraints" (TOC) will be presented, which preaches that the management of any
value adding system should pay special attention to the “bottleneck” resources or the
resources with lowest capacity. The fundamental principles of the TOC will be
presented and how to apply them to practical situations will be discussed.

15.2 CONCEPTS
15.2.1 WHY “THEORY OF CONSTRAINTS”?
Some researchers working in Israel in the 1960s and 1970s, trying to build a
computer simulator of pulled production systems (lean style) realized that the lowest
capacity resources in any system often constrained the capacity of the entire system.
They then began to develop an alternative logic to that prevailing at the time (of looking
at the management of operations as if the capacity of all resources was equally relevant
in the treatment of the system’s capacity). The evolution of these ideas led to the
creation of the so-called "Theory of Constraints" (which considers “constraining”
capacity resources to be especially important, not only for capacity management, but
for various other aspects of operations management as well). The Theory of Constraints
has been popularized by the writings of one of its creators, the late physicist turned
business consultant and guru Eliyahu Goldratt. Let's look at how this theory applies to
issues related to production and operations management.

15.2.2 OPTIMIZED PRODUCTION TECHNOLOGY (OPT)

The ideas that originated the Theory of Constraints, were initially known by the
acronym OPT. OPT stands for "Optimized Production Technology", a production and
operations management technique developed by a group of Israeli researchers, which
included the physicist Eliyahu Goldratt. Going forward, the text sometimes refers to
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

the Theory of Constraints as OPT/TOC to emphasize that OPT is a production

scheduling system intrinsically connected with the Theory of Constraints. It is
important to clarify, first, that, although the name by which the technique became
known suggests that it “optimizes” production this is not strictly true because
OPT/TOC is not an optimizing technique in the scientific sense of the term. There is
no guarantee that optimal solutions are reached by its application since the technique
is based on a series of heuristic procedures (these are systematized “common sense”
algorithms).

Objectives
The OPT/TOC approach advocates that the basic goal of companies is to “make
money”. It also considers that manufacturing should contribute to this basic objective
by acting on three elements: increasing Throughput, reducing Inventories and reducing
Operating expenses. These terms must be defined in such a way as to avoid confusion
with their more common semantic meanings. According to the OPT/TOC approach:

● Throughput: is the rate at which the system makes money by selling its products.
It should be noted that throughput refers to the flow of products that are sold.
Products made but not sold are still considered only as “inventory”.
● Inventory: is the money that the company spent in the goods it intends to
transform and sell. It refers to the value of the raw materials and components
involved only. You don't include “added value” or “labor content” here. The
traditional “value added” by labor is included in the Operating expenses.
● Operating expenses: the money the system spends to turn Inventory into
Throughput.

OPT/TOC proponents argue that if a company simultaneously achieves the

objectives of increasing throughput, reducing inventory, and reducing operating
expenses, it will also automatically improve its performance in the goals of increasing
net income, return on investment and cash flow (measures most traditionally accepted
in the corporate financial world). This correspondence between sets of performance
measures is explored by Goldratt in his famous book “The Haystack Syndrome”.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

There would be advantages of adopting the objectives throughput, inventory, and

operating expenses instead of traditional net income, return on investment and cash
flow: it would be easier for people linked to the operation to associate their actions and
decisions with the new objectives than to associate them with the traditional ones, thus
being able to make better decisions, towards achieving the above goal: “making
money” for the company. For example, a machine operator understands what they can
do in their day-to-day life to “reduce inventories” but has more difficulty imagining
what can be done by them to increase "return on investment", a much more abstract
concept for the operator.
To maximize the achievement of these objectives, OPT/TOC questions and
challenges some assumptions that traditional production management has considered
as postulates, especially in relation to the activity scheduling aspect. Basically,
OPT/TOC considers that there are four areas – Resource Types, Machine Preparation,
Lot Size and The Effects of Uncertainties – that deserve to be reconsidered. These four
areas will be discussed below in this chapter.
Despite the complex nature of the bottleneck concept and its developments in the
theory of constraints, its logic is very important, especially for companies that have
their operations strategy based on delivery speed and reliability. Imagine, for example,
the recent cases of Wal-Mart and Amazon.com, companies that have developed
different delivery skills. Both deal with ever-increasing volumes of product processing
for delivery and, on top of this, have also developed operations that allow the delivery
of many products on the same day or the day after purchase by customers. These
qualities depend on a series of operations management actions and processes that must
carefully consider the concepts of Theory of Constraints.

Did you know?

Eli Goldratt and his colleagues began their studies that resulted in The Theory of
Constraints while working as professors in a physics department at a University in
Israel. They tried to model “pulled” production systems (such as in JIT/lean) on a
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

computer. They soon noticed that in pulled systems the lowest capacity resources
naturally limited the flow. This drew their attention to the idea of treating
“bottlenecks” (capacity constraining resources) especially.

15.2.3 IMPORTANT CONCEPTS IN THEORY OF CONSTRAINTS

Resource Types
To schedule activities properly, to allow the achievement of the aforementioned TOC
objectives, OPT/TOC considers that it is first necessary to understand very well the
interrelationship of the two types of resources that are normally present in all
operations systems:

● Bottlenecks – capacity constraining resources.

● Non-bottlenecks – non-capacity constraining resources.

The resources can be understood here as any element necessary to produce a product,
such as people, equipment, devices, measuring instruments, space, etc. Consider a
resource X and take it that the total availability of X is 200 hours per month. The
potential market demand reflects a use of that resource for more than 200 hours per
month. Resource X therefore is a bottleneck – a resource with less availability than the
market demand would potentially require. As resource X only has 200 hours available,
all of them end up being utilized to serve (albeit partially) the market. Therefore, by
definition, a bottleneck resource is busy for its entire available time. Now consider
another resource Y, a non-bottleneck and probably a relatively faster resource. All
market demands would only require it to be busy 150 hours a month. It also, like
resource X, has a total availability of 200 hours per month. So by definition, a non-
bottleneck resource has more availability than all demand would require.

Did you know?

Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

Bottleneck resources, by definition, do not spend any time being idle; non-
bottlenecks, also by definition, spend at least some time being idle. This doesn’t
seem very important, but the implications of this are what originated the TOC.

The nine principles of the theory of constraints in operations

Let's now look at four possible types of relationships between these two types of
resources – bottleneck (X) and non-bottleneck (Y) resources (see Figure 15.1).

FIGURE 15.1 RELATIONSHIPS BETWEEN BOTTLENECK AND NON-BOTTLENECK RESOURCES.

To reflect
If a manager asks you to look at the 4 cases described in Figure 15.1 and she asks
you to maximize the “utilization of all resources” to 100% in each case (cases 1, 2,
3, and 4), what would your answer be? (for each case):

1. Yes, we should do this because it is possible, and it is the best thing to do.
2. Yes, we can do this, but we shouldn't do it.
3. No, it is impossible to do this.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

● CASE 1: When production flows from resource X to resource Y. In this situation,

resource X (100%) should be fully utilized, but resource Y can only be utilized
75% of the time. Resource X, because it is a bottleneck resource, cannot produce
enough to keep resource Y working all the time. Therefore, your answer should
be “No, it is impossible to do this” in case 1.
● CASE 2: When production flows from Y to X. Again, resource X is utilized 100%
of the time, and if there is enough raw material, you can activate resource Y
100% of the time as well. However, remembering that one of the objectives of
OPT/TOC is to simultaneously increase throughput and reduce inventory and
operating expenses, we should conclude that Y should only be activated 75% of
the time, because by activating Y more than 75% of the time would cause
inventory to buildup in the process between resource Y and resource X, without
any increase in throughput, which is limited by the bottleneck (X) capacity. In
this situation, from the perspective of OPT/TOC, if resource Y is utilized for
more than 75% of the time, this action would be contrary to achieving OPT/TOC
objectives. Your answer should be “Yes, we can do this, but we shouldn't do it”.
● CASE 3: Resources X and Y, instead of feeding on each other, feed an assembly
operation that uses a processed part (one unit) of each. Again, resource X should
be utilized 100% of the time to maximize throughput. However, if resource Y is
activated for more than 75% of the time, inventory accumulates before assembly,
as the overall flow will also be limited by the production capacity of bottleneck
resource X. Therefore, resource Y should, in this situation also, according to
OPT/TOC, be activated only in 75% of its available time. The answer again
should be “Yes, we can do this, but we shouldn't do it”.
● CASE 4: Resources X and Y do not feed each other or feed a common assembly,
but supply independent market demands. Once again, resource X should be used
100% of the time, but resource Y should only be used 75% of the time, under
penalty of building up stocks of finished products, as demand remains limited
and, to meet it, the use of resource Y for only 75% (150 hours per month) of the
time is sufficient. The answer again should be “Yes, we can do this, but we
shouldn't do it”.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

The many other forms of relationship between bottleneck and non-bottleneck

resources can normally be described by combinations of the four previous simple
situations.
Now that we introduced the topic, let us describe the nine principles of OPT/TOC.
Traditionally, especially in mass production, the idea was to balance the capacity of
the various stages of a production process and then try to establish a smooth material
flow, if possible, a continuous one.
OPT/TOC advocates that capacity balancing is impossible to achieve in most
situations that deviate from the strict “mass production” conditions, as with an
increasing variety of products being processed in the production unit, the relative
productivity of the different product mixes will fatally result in unbalances. Therefore,
if it is a fact that the production units will mostly have unbalanced capacities, what
should be sought is a balanced production flow in the factory, ensuring that, at each
stage of a production process, the flow that goes through the system is balanced. This
is a way to prevent unwanted in-process inventories from building up in the system –
which will fatally arise whenever in a process the flow of one step overcomes the flow
of the next step.

Principle 1: Balance flow, not capacity.

As shown in Figure 15.1, the use of a non-bottleneck resource should not be defined
by its own availability, but by the capacity of the system’s bottleneck resource. In cases
1, 2 and 3, for example, the utilization of a non-bottleneck resource such as Y should
be determined by the capacity of the bottleneck resource X. In case 4, it should be
determined by market demand, which is the bottleneck in that case. However intuitive
this may seem, it is also interesting to note that in the practice of companies, this
principle is often not considered or overlooked. It is very common to find performance
evaluation systems in manufacturing operations that consider as central metric the so-
called "resource utilization" indexes (some variation of the percentage of time that the
resource spends busy). This regardless of whether the resource in analysis be a
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

bottleneck resource or not. Imagine this being done with our case 2 of Figure 15.1. The
result would be that the operations manager, in the process of maximizing its “resource
utilization” performance metric, would seek to keep the non-bottleneck resource Y
100% of the time activated! It would create inventory without any increase in the
system’s throughput, thus going against the organization’s goal of “making money”.

Principle 2: The utilization of a non-bottleneck resource should NOT be determined

by its own availability, but by the capacity of some other system constraint (for
example, a bottleneck).

From OPT/TOC's perspective, there are important distinctions to make between

activating and utilizing a resource. Activating a non-bottleneck more than the time that
is sufficient to feed a bottleneck resource does not contribute at all to the objectives
defined by OPT/TOC. On the contrary, the throughput would remain constant, still
limited by the bottleneck resource and, at the same time, inventory would increase and
probably also operating expenses would increase, with the administrative expenses
resulting from the generated inventory. Because the extra activation of the non-
bottleneck resource does not imply a contribution to the achievement of the TOC
objectives, it cannot be called “utilization” of said resource (since nothing “useful” is
being generated) – it is only a mere “activation” of the resource.

Principle 3: Utilization and activation of a resource are not synonymous.

Machine setup
To understand the following principles, it is important that we analyze bottleneck and
non-bottleneck resources, but doing so from another point of view: what do the two
types of resources do with the time they have available? Look at Figure 15.2. During
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

its available time, by definition, the bottleneck resource is either being used to
effectively process flows (of materials, information, or people) or is being prepared
(being “setup”) to process the next product. If there was a third type of time, it would
be “idleness” – but if idleness is found in a resource, by definition the resource will be
reclassified to a “non-bottleneck.”

FIGURE 15.2 COMPONENTS OF THE AVAILABLE TIME OF THE TWO RESOURCE TYPES.

A non-bottleneck resource on its turn spends its available time doing one of three
possible activities: either it is processing its flows, or it is being setup to process its
flows, or it is idle. This is because if there is no idleness, this indicates that the resource
in question is a bottleneck, not a non-bottleneck resource.
In a bottleneck resource, if an hour of preparation time is saved, an hour is
automatically gained to be used in processing, that is, the bottleneck resource gains an
hour of availability to process material. In addition, an hour gained for processing on a
bottleneck resource is not merely an hour gained on the bottleneck resource, but an
additional hour gained across the entire production system, since it is the bottleneck
resource that limits the overall system’s capacity to make money. The OPT/TOC seeks,
whenever possible, to keep the production batches as large as possible in the bottleneck
resources, exactly to minimize the number of setups to be performed in the bottleneck
and, therefore, minimize the total time spent on setting up such resources, consequently
increasing the ability of the bottleneck to process more flow and thus increasing the
throughput of the entire system.

Principle 4: An hour gained in a bottleneck resource is an hour gained for the entire
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

system.

To reflect
In a fast-food restaurant like McDonald's, what do you think is the bottleneck
resource? If the bottleneck resource constrains the entire corporation's capacity in
terms of how much money they make, what could you do to reduce the bottleneck
idleness to a minimum during peak demand hours?

Note, in Figure 15.2, the available time components of a non-bottleneck resource.

Setup time saved on a non-bottleneck resource is only an hour idle longer for that
resource, since the processing time of a non-bottleneck resource is determined, not by
its own availability, but by the bottleneck’s capacity.

Principle 5: An hour gained on a non-bottleneck resource is nothing, it’s just a

mirage.

When scheduling resources, therefore, in operations involving a bottleneck resource,

it is important to save time with setups, both by reducing the time spent on each setup
(via quicker changeovers) and by reducing the total number of setups (or, in other
words, by processing larger batches), thus allowing throughput to increase. However,
in an operation involving a non-bottleneck resource, there is no obvious benefit of
reducing setup times. In fact, there would even be the convenience of using part of the
idle time to make a greater number of setups, because, in this way, the lot sizes and
therefore the average inventory of the system would be lower, and the flow speed would
increase. While not necessarily smaller batches help increase throughput, they do help
decrease in-process inventory and operating expenses, making production flow
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

smoother as smaller batches flow faster through the production unit.

Lot size
According to the OPT/TOC philosophy, the two previously discussed principles
suggest that the calculation of the economic lot (batch) size should not be applied in
the traditional way. The traditional economic lot size calculation considers that annual
setup costs decline as the processed batch size increases and that this would be valid
for all resources. See Figure 15.3. This would not always be valid tough, if it is true
that an hour gained on a non-bottleneck resource does not represent the same as an
hour gained on a bottleneck resource. Gaining an hour on a bottleneck resource is not
only gaining an hour from a machine setup team or an hour of production on the
machine, but it also means gaining an hour for the whole system. Similarly, an hour
gained on a non-bottleneck resource is a mirage.
Another point suggested by OPT/TOC as important regarding production batch sizes
is the difference between the batch sizes seen from the material flow point of view and
from the resource point of view. OPT/TOC advocates often use an example to explain
this point: imagine a continuous, dedicated production line. What is the batch size of a
production flow on a continuous production line? Two types of answer are possible:
one is that the batch size is 1, because the products are moved one by one, from one
workstation to the next. Another possible answer is that the batch size is “infinite” (or
very large) because the line is dedicated, and a very large number of products will be
produced before the line is stopped so that it is setup for the production of a different
product.

To reflect
Isn’t it interesting that traditional batch size definition techniques, such as the logic
of the “economic lot size” ignore in their formulation whether the resource they are
made in is a bottleneck resource or not? Why do you think that is?

In principle, the two answers could be considered correct, depending on the point of
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

view that is considered. From the point of view of the flow or, for example, of a printer
being assembled, the batch size is 1, since the units are passed one-by-one along the
workstations. From the point of view of the resource or, for example, of a workstation,
the batch is very large, since for a long time this station will be setup for the same type
of product to be processed. According to the OPT/TOC, it is necessary to consider the
issue of lot sizes according to these two perspectives:

FIGURE 15.3 THE LOGIC OF THE ECONOMIC LOT AND THE OPT/TOC.

● The perspective of the resource: related to what is called, in OPT/TOC,

“processing batch”.
● The perspective of the flow: related to what is called, in OPT/TOC, “transfer
batch”.

Principle 6: In a process, the transfer batch may not and should often not be the
same as the processing batch.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

In OPT/TOC, the transfer batch is always a fraction of the processing batch. The
processing batch is the batch size that will be processed on a resource before it is re-
setup for processing another item. The definition of transfer batch is the size of the
batches that will be transferred to the next operation. As in OPT/TOC these batches do
not necessarily have to be the same, quantities of processed material can be transferred
to a subsequent operation even before the entire processing batch is processed. This
allows the batches to be divided and can substantially reduce the throughput time of
the products in the factory. See Figure 15.4.

FIGURE 15.4 TRANSFER AND PROCESSING BATCHES.

Unlike OPT/TOC, many traditional production scheduling systems assume that

there is only one batch size for each item, that is, they consider that the transfer batch
is always the same as the processing batch. Some traditional systems also consider that
this batch size should be the same for all operations that process the product. This
brings an additional difficulty to the determination of lot sizes, since, for a process that
requires the processing of several machines sequentially, it is possible that for each
machine the calculation of the economic order size results in a distinct result. Which,
then, do you adopt? In OPT/TOC systems, this constraint does not exist, and it is
possible for the batch size to vary throughout the production process of a given product.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

This leads to another OPT/TOC principle:

Principle 7: The processing batch must be variable, not fixed.

In OPT/TOC, unlike in most traditional systems, the size of the processing batches
is a function of the factory situation and can potentially vary from operation to
operation. These batch sizes are established by the OPT/TOC calculation system,
which considers the costs of carrying inventories, setup costs, flow requirements for
certain items, resource types (bottleneck or non-bottleneck), and more.

Effects of uncertainties
Uncertain events (events with random variability) will fatally occur in a complex
system, such as production systems. As it is extremely difficult to anticipate where, in
the system, uncertain events will occur, it is necessary that the system is protected at
its fragile or critical points. In addition, the production of an item typically involves
various material processing and transportation operations.
For most of these operations, the processing time varies according to statistical
distributions. In other words, the processing time of the same operation varies
randomly according to some distribution each time the operation is performed. This
means that in production planning and scheduling, when using processing times or lead
times for a certain operation, in fact, the averages, or expected value, of processing
times or lead times, are actually what is being considered. But they are subject to
statistical fluctuations.
Such fluctuations may be due to uncertainties in the operation, lack of operator
consistency, limits on equipment capability, equipment breakdowns, among others. As
much of this statistical fluctuation can be controlled to a certain extent, via operator
training, standardization of work methods, automation of tasks, better preventive
maintenance, among others, it is usually impossible for production systems to
completely eliminate the random component of the processing times of their
operations. Therefore, virtually in all production processes statistical fluctuations exist,
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

to a greater or lesser extent, and they potentially affect a large part of the operations
performed.
Statistical fluctuations can have an approximately normal distribution since they are
the result of the occurrence of a series of random or out-of-control events. If
manufacturing operations were isolated, i.e., if they were not part of a sequential chain
of operations that contribute to the production of a given item, these statistical
fluctuations would tend to cancel out and tend to be close to zero in cumulative terms.
Simply put, delays in certain cycles would tend to offset advances in others, so that, on
average, the deviation from the expected processing time of the operation would tend
to zero. However, in fact, manufacturing involves chaining interdependent operations,
that is, typically a given operation can only be performed when the previous operation
in the chain ends. Therefore, in this case, the statistical fluctuation of the chain tends
not to cancel out, but delays tend to propagate along the chain. A five-minute delay in
a chain operation causes the subsequent operation to only begin five minutes later,
because the two chained operations are dependent events. The combined effect of
statistical fluctuations and chaining of dependent events can be illustrated by an
example shown in Figure 15.5.1
In this example, there are two resources (for example, workers operating machines),
A and B. The performance of resource A varies according to the normal distribution
shown. It, on average, takes 10 hours to process each piece, but times between 8 and
12 hours are considered within the acceptable "normal" range.

1
This example, as well as some of the others, regarding the operation of the OPT/TOC were originally developed by John Helliwell.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

FIGURE 15.5 PROPAGATION OF THE EFFECT OF STATISTICAL FLUCTUATIONS IN

DEPENDENT EVENTS.

Worker B operates an automated machine, whose performance is highly consistent,

with an average of 10 hours and very small variation. It produces each piece in 10 hours
(almost exactly). In this example, all production flows from worker A to worker B. The
production schedule initially defined the production of four items and shows worker A
working on item 1 from instant 0 to instant 10 hours; on item 2 from instant 10 to
instant 20 hours, on item 3 from 20 to 30 hours, and on item 4 in period 30 to 40. As
there is no initial inventory in this hypothetical problem, the schedule of worker B
shows that he should work on item 1 in period 10 to 20; on item 2, from 20 to 30; on
item 3, from 30 to 40, and finally, on item 4 from 40 to 50 hours. This schedule seems
logical.
Let's now look at what really happens at the factory. Consider that Worker A had a
difficult weekend, and when he came to work on Monday morning, he wasn’t on his
most efficient days. So, it took him 12 hours to finish the first item. At that point, when
he finished the first product, he was two hours behind schedule. His state of relative
(although expected) inefficiency continued during the execution of the second item and
he, to finish it, had to work from instant 12 to instant 24, when it ended. Now, worker
A is four hours behind schedule.
As the example company has an agile control system, the line supervisor realized
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

that worker A was four hours behind and asked him to speed up. As worker A is very
aware of his responsibilities, he struggled and finished the third item in eight hours (24
to 32). At that point, he was only two hours behind the schedule. In an extra effort, he
also completed item 4 in eight hours, ending the week meeting his individual schedule.
Let's now see what happened to worker B. The schedule stated that he should begin
his work on item 1 at the end of instant 10, but, as there was a delay in the previous
stage, he could not start before 12. He then worked on item 1 from 12 to 22. At that
point, he was two hours behind schedule, despite having worked exactly as expected,
in 10 hours. The delay of worker A had been propagated and reached worker B. At
instant 22, then, worker B is ready to start working on item 2, but he still must wait for
the material, available only at instant 24.
Consequently, he works on item 2 from instant 24 to 34. The material in item 3 is
ready to be worked on instant 32, but worker B cannot start working on item 3 because
he is still working on item 2. He then completes item 3 at instant 44 and finally
completes item 4 at instant 54. Even though he performed exactly as expected in the
production of the four items, he finishes the work four hours behind schedule and the
line as a whole endures a lengthy delay. The conclusion is simple: with dependent
events, statistical fluctuation does not tend to be cancelled out in the long run, but to
accumulate. Worker B finished his work delayed four hours, although worker A
finished his work within schedule. Once the phenomenon is identified, which is one of
the main causes of delays in the factory, it is now important to also identify ways to
establish actions to contain and isolate the damage.
If worker B was a system bottleneck, the whole system would have lost, never to
recover, four hours of flow, waiting for the delayed item of worker A operation.
Considering, then, that both statistical fluctuation of uncertain events (e.g., machine
breakdowns) together with the existence of dependent events can seriously
compromise schedule compliance (and even compromise the utilization of bottleneck
resources), it is important that the system is protected, especially in terms of its most
critical resources, i.e., their bottlenecks resources.
This finding is one of the motivators of another OPT/TOC principle:
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

Principle 8: Bottlenecks not only determine the flow of the system, but also define
its inventories.

Did you know?

Principle 8 occurs because the phenomenon in analysis is asymmetric. When a
resource has queue waiting for processing, inventory accumulates. When idle, the
“wasted capacity” does not accumulate to be used later.

Bottlenecks define the flow of the production system because they are the capacity
constraining resources, as commented earlier. However, they are also important
determinants of inventory, as stocks are sized and located at such points that they can
isolate bottlenecks from statistical fluctuations propagated by non-bottleneck resources
that feed them. For example, a stock is built before the bottleneck machine, so that any
delay (whether caused by statistical fluctuation or random events) does not cause
bottleneck starvation (stoppage for lack of material).
Typically, this is done by creating a time buffer before the bottleneck resource. In
other words, the materials are scheduled to arrive at the bottleneck resource a certain
(safety) time before the moment this resource is scheduled to start operating on them.
This way, if any delay occurs with the resources that feed the bottleneck resource, this
can be absorbed by the safety stock.
An example of the use of a time buffer in service operations is what many doctors
do in their offices. In an attempt to reduce the uncertainties of the process of arrival of
patients, which would compromise the efficient use of their time or the waiting time of
patients, they establish a system of appointments (if left free to decide the time to go
to the doctor, in many situations the arrival of patients would be mostly on the time of
departure from work thereby increasing queues and waiting times and, at the same
time, at times such as mid-afternoon, the doctor could stay idle). With appointments,
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

the “arrival rates” of patients become more uniform, with less “statistical fluctuation”.
However, customers, even with appointments, are still subject to uncertainties, for
example, of traffic. A delay of a patient can cause the doctor to stay idle. What some
doctors do, then, is schedule more patients (a type of overbooking) than their strict
ability can meet in a period. This means that even when scheduling an appointment,
the probability is that the patient, even arriving on time, still waits a while for service.
This, from the doctor's point of view, means that if there is some delay of the patients,
this will not mean idleness. Therefore, there is a “time buffer” of patients, who protect
the bottleneck resource, which, in this case, is the doctor’s scarce time.

To reflect
What disadvantages do you identify in the use of overbooking to minimize the
idleness of doctors in their offices?

Now, think about other situations in service operation where you noticed that the
principles of TOC are used, describing each of them.

Lead times and priorities

Traditional MRP systems used to plan and schedule production and materials are based
on the assumption that lead times can be determined a priori of the planning process.
In fact, MRP algorithm requires lead times as input data for the planning process.
Starting from the promised delivery date, MRP offsets the lead times for the various
product components, to reach the dates of production and purchase of such components
(this logic is known as “backward scheduling”). These lead times are, in turn,
dependent on the estimated queue times awaiting processing at workstations. Once the
priorities have been established by the MRP (based mainly on the said lead times), the
system will then check if there is enough capacity in the system to comply with the
program. Priorities or, in other words, scheduling and capacity are considered
sequentially and not simultaneously in MRP systems – which first schedules activities
based on the lead times of materials and then checks whether there is sufficient capacity
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

to perform said activities.

OPT/TOC, conversely, considers that queue times are dependent on how scheduling
is done. In fact, if a given production order gains priority for any reason in a queue of
orders waiting for processing, that order will remain a shorter time in the queue. As
queue time is one of the main components of the lead times of items, it is evident that
lead times will be different, depending on how the orders are prioritized, sequenced
and scheduled. However, if lead times are a result of the scheduling process, they
should not be used as an input data for that process – which is exactly what MRP does.
The OPT/TOC approaches the problem differently, considering the scheduling of
activities and the capacity of resources simultaneously, especially activities performed
in bottlenecks. Considering the capacity limitations of the bottleneck resources, the
OPT/TOC system then decides for priorities in the occupation of these resources and
based on the defined sequence of operations that occur in the bottleneck, it can
calculate, as a result, the lead times much more precisely and, therefore, can more
appropriately schedule production. This can be substantiated by the ninth opt/TOC
principle:

Principle 9: Activity scheduling and production capacity should be considered

simultaneously and not sequentially. Lead times are a result of scheduling and cannot
be assumed a priori.

OPT/TOC advocates argue that this feature makes OPT/TOC-generated schedules

more realistic than those generated by MRP.

How OPT works

Drum-buffer-rope

One of the strongest points considered in the OPT/TOC system refers to the way it
schedules activities. The OPT/TOC scheduling is based on the nine principles
previously presented. The main aspects of this scheduling system are described below.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

In a manufacturing environment, there are several constraints to be considered:

market, supplies, and the capacity of the production process itself are some examples.
There may not be real bottlenecks, but there will always be Capacity Constraining
Resources (CCR). CCR, like bottlenecks, control the flow and must be synchronized
with other resources in order to be able to control inventories.

Did you know?

The difference between a bottleneck and a CCR is simple. In some situations, there
may be no real bottlenecks in a factory – all production centers are oversized relative
to demand; however, there will always be some resource that constrains production.
This, then, will be the CCR, although it is not a real bottleneck. There may also be a
case where, by definition, multiple resources are bottlenecks (several resources have
less capacity than the market demand). In this case, the CCR will be that resource,
among those, that has the least production capacity. This (CCR) will then limit the
production capacity of the entire system. In this text, except when explicitly stated,
the terms "CCR" and "bottleneck" will be used indistinctly.

If a simple route composed of nine resources is considered and it is discovered that

resource number 6 is a CCR or bottleneck, the OPT/TOC system will start scheduling
at exactly that point. This is the point at which it is necessary to “beat the drum”, which
is the analogy used by the creators of OPT/TOC to symbolize that this is the point that
should dictate the rhythm of the entire production system.
First, OPT/TOC fully loads the constraining resource (CCR) to achieve maximum
flow. At the same time, the best sequence for the work is established, deciding the
priorities between the activities, considering the dates of the requested orders.
Secondly, the CCR should be protected against possible uncertainties that could
jeopardize the arrival of materials for the fulfilment of the working sequence that the
CCR will perform. If it is found that an uncertain event with considerable probability
of occurring with the supply of material to the CCR is for instance the breakdown of
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

the supplying machine, with an expected repair time of two days, it is necessary to plan
the arrival of the materials coming from that supplying machine to the CCR at least
two days before the date on which the CCR is scheduled to process them.
This stock is called “time buffer” because materials passing through this stocking
point are constantly changing, in this case, they will have value of at least two days of
bottleneck processing. Third, it is necessary to use the CCR schedule to schedule and
control the use of non-bottleneck resources.
The non-bottlenecked resources that come, in the production route, after the
bottleneck resource are controlled directly by the CCR, since they can only process
what is released by the CCR and in the sequence in which the CCR releases them. By
definition, non-bottleneck resources have slack (idleness) in the schedule, that is, they
have greater production capacity than what is demanded from them. Therefore, they
should have no problem processing the material coming from the bottlenecks.
A CCR also controls inventories throughout the production process. This is done by
"tying a rope" that connects the stock created by the “time buffer” to the initial
operation of the production system. Thus, the first operation is only scheduled, or, in
other words, raw materials are admitted to the system to begin to be processed, in a
synchronized way and according to the future needs (in the precise quantities and at
the right time) of material arrival in the protective stocks (the time buffers) of the CCR.
Thus, stocks will not be able to build up to higher levels than the predetermined
ones, considering the possible uncertainties to which non-bottleneck resources prior to
CRRs may be subject (remembering that the idea of time buffers is to ensure that CCR
is never idle).
The synchronization described is named in the literature on OPT/TOC drum-buffer-
rope, in reference to the trio of elements that are keys to the method: drum-safety time
buffer-rope. The drum, representing the CCR, dictates the rhythm, flow and volume of
the production of the whole system. The protective stock, defined as a safety time
buffer before the CCR and synchronized with it, ensures that the CCR does not stop
for lack of material, and finally, the "rope", which represents the synchronization
between the need for arrival of materials in the safety time buffer and the admission of
raw materials in the system.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

OPT/TOC seeks to put the drum-buffer-rope idea into practice by using a finite
scheduling program forward (loading orders on CCR resources sequentially in time,
considering their capacity constraints), to try to maximize the flow processed by the
CCR. The scheduling sequence follows an algorithm that the company that hold
OPT/TOC rights do not make public. However, they report that this algorithm
considers:

● The delivery due dates for orders.

● The convenience of reducing the number of machine setups.
● The factory “congestion” situation.
● Secondary time buffers.

Consider a factory that has a final assembly operation. Also consider that the CCR
of this plant is in one of the branches that feed this assembly, as shown in Figure 15.6:

FIGURA 15.6 THE DRUM-BUFFER-ROPE LOGIC IN OPT/TOC.

Considering the situation in Figure 15.6, it is also necessary to provide the branch
that does not contain the CCR with some kind of protection, because, without this, the
parts already processed by the CCR could have to wait to be assembled, if the parts
provided by the branch with no bottleneck suffer some kind of delay.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

As one of the objectives of OPT/TOC is to increase throughput, and this is defined

as the flow of material sold by the system, OPT/TOC seeks to accelerate the completion
of orders as much as possible, after they are processed by the CCR. Thus, OPT/TOC
defines some secondary time buffers (in addition to those described, which aim to
protect the CCR themselves), which protect the promised delivery dates and the
throughput of the system. These secondary time buffers are made up of materials that
must participate in assembly along with material that has already been processed by an
CCR.
The goal is to ensure that materials processed by CCR can, as soon as possible,
become throughput (be sold), not being subject to uncertain events of non-bottleneck
operations that may occur and that could potentially affect the maximized flow
intended by the system. The size of the secondary time buffers is defined by the nature
and probability of the occurrence of random events that can affect the performance of
non-bottleneck branches.
OPT/TOC combines two scheduling logics – finite forward scheduling and infinite
backward scheduling, respectively to schedule bottlenecks and non-bottlenecks. This
can be better understood through the diagram in Figure 15.6.
Component A1 is initially processed in the non-bottleneck resource Y1, and then in
the CCR resource X. It is thus processed in two other non-bottleneck resources Y2 and
Y3. The A2 component is processed on the non-bottleneck resources Y4 and Y5.
CCR X also processes other items: L, M, N, B, C, D, E, and others. The working
sequence in the CCR was decided to be L-M-N-A1-B-C-D-E, that is, programming
forward and sequentially occupying the capacity of resource X in time. From there, the
beginning and end moments of the processing of component A1 in CCR X are defined.
By setting the instant that processing A1 should begin in resource X, a time buffer is
established before resource X, to protect it from uncertainties of the branch that
processes A1. This means programming the arrival of A1 in stock before resource X,
for a certain time before the time scheduled for the start of processing.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

FIGURE 15.7 OPT/TOC BACKWARD AND FORWARD SCHEDULING LOGIC.

The OPT/TOC then schedules, from the moment it is necessary that material A1
reaches the protective time buffer, the moments when the Y1 operation must begin and
end and the purchase of raw materials for A1 (scheduling backward in time).
Based on the end-of-processing instant of A1 in resource X, we can calculate by
adding the processing times in resources after resource X (Y3, Y4, and the final
assembly), the instant you expect to finish processing order A. But it is not only
necessary for component A1 to be available for final assembly. It is also necessary that
component A2 be available at the time of the start of the final assembly.
In order not to run the risk that component A1, coming from resource X (the CCR),
is still waiting for component A2 to arrive for assembly (something unexpected may
happen in the non-bottleneck branch), OPT/TOC schedules component A2 to be
completed and be available for final assembly a certain period before the expected time
for arrival in that assembly, of component A1.
From the definition of the instant when component A2 should be available for final
assembly, OPT/TOC calculates, through backward scheduling logic, what are the times
when Y4 and Y5 operations and the purchase of raw materials for A2 should begin and
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

end.
In this way, the forward and backward scheduling logics are combined to generate a
complete schedule for the production system. It is important to note that it is essential
to know precisely what are the bottlenecks of the production system that is intended to
be managed with OPT/TOC. Otherwise, the entire scheduling will be based on false
premises, compromising the quality of the OPT/TOC solution.

OPT decision-making process

The creators of OPT/TOC suggest that the following process be followed to manage a
Production system that uses OPT/TOC:

● Step 1: identifying the capacity constraint(s) (CCRs or bottlenecks) – identify

resources whose production capacity constrains the system’s capacity to meet its
product sales flow. Attention to the fact that it is possible that the constraint is
in the market itself.
● Step 2: exploit the constraint(s) of the process – exploiting the constraints simply
means making the most of them. Do not waste any time on bottleneck machines,
for example. Using constraints as much as possible is what this step means.
● Step 3: subordinate all other decisions to the identified constraints – bottlenecks
define production flow and inventories, the utilization of non-bottleneck
resources, among others.
● Step 4: try to elevate the constraint – it means somehow increasing the
production capacity of the bottleneck to increase the capacity of the system. This
step should only be taken after the constraint has been exploited to the fullest, as
it may have an impact on operating expenses (e.g., with subcontracting, extra
shifts, purchase of machines, etc.).
● Step 5: If in step 4 a constraint has been elevated, go back to step 1 to identify
the next system constraint.

INDUSTRY 4.0: THEORY OF CONSTRAINTS IN OPERATIONS

Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

Currently, the application of information systems and software such as

Advanced Planning Systems (APS) contribute widely to achieve the objectives
of OPT/TOC and many make use of its principles, by synchronizing supply and
demand to schedule production and prepare inventory plans, maximize
production in bottleneck resources to increase earnings, create optimized
schedules in order to enhance production efficiency and reduce operational
expenses, and thus enable decision-making based on current data.

However, the support of systems in manufacturing operations is increasingly

limited when facing varied and increasingly complex production environments,
still considering that regardless of the processes of technological and digital
transformation, there will always be constraints in the processes of any type of
organization, even in fully automated factories, provided that equipment and
machinery, at some point in production, they will end up being bottlenecks and,
therefore, candidates for the application of TOC.

In the context of I4.0, TOC is mainly incorporated into real-time communication

between the components of production lines to optimize throughput. In more
efficient scenarios, warnings issued by smart machines and lean tools will
anticipate faulty operation that could lead to any kind of jeopardy to bottlenecks.
Consequently, these tools allow proactive resource planning to effectively
manage constraints and achieve higher throughput.

In this perspective, with the support of big data analysis for data exploration and
machine learning as a branch of artificial intelligence, the development and
application of predictive algorithms oriented to the prediction of bottlenecks in
complex production systems is increasingly used. These algorithms are tested in
simulation environments to accurately mimic real systems, so simulation allows
the measurement of equipment performance and performance on production
lines, revealing possible constraints during the process. Thus, simulation
environments as a validation tool can be combined with TOC.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

With Flexsim software, for example, it is possible to create discrete simulation

models based on Virtual Reality and Digital Twin environments in 3D processes
that allow you to identify system constraints, evaluate the effectiveness of
available alternatives and predicting possible shifting of the constraint.

The following video exemplifies the use of simulation and TOC in operations
management: https://youtu.be/zxTLqMg0SW8. Access on Feb. 12. 2022.

15.3 CASE STUDY

Imagine yourself in the following situation: you must support a decision maker on
whether to purchase the Lucas Risso Ltd. factory. The products made by Risso and its
production processes are represented in Figure 15.8.
The production unit in question manufactures two products, P and Q.
Inputs: product P is made from two raw materials, RM1 and RM2, and one
purchased item; one unit of each is required to produce one unit of product P. Product
Q is made from two raw materials, RM2 and RM3, does not use the purchased item
and also requires one unit of each for a unit of product Q to be produced. Raw materials
all have a unit cost of $20, and the purchased item has a cost of $5 per unit. Suppliers
are reliable in terms of quality, quantity and time (delivery is always immediate,
exactly in the necessary quantities and with perfect quality).
Process: There are four types of resources available to process the inputs: resources
A, B, C, and D. They are not interchangeable (each resource type does not perform other
resources’ activities) and only one unit of each resource is available for use during one
shift (eight hours), five days a week. Preventive maintenance is perfect, and resources
are available for 100% of the working period. The processes necessary for the
processing of each of the raw materials and the purchased item are represented in
Figure 15.8. RM1, for example, must be processed in resource A for 15 minutes (per
unit) and in resource C for 10 minutes (per unit) in any sequence. The result is a “semi-
finished” item that, along with the purchased item and another semi-finished item
(result of the processing of RM2 for 15 minutes per unit in resource B and 5 minutes
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

per unit in resource C), is processed for 15 minutes (per unit) in resource D – resulting
in final product P. Product Q has a process also described in Figure 15.8. There is no
time or cost of setup, when changing the processed product in the resource. The quality
of the processes is impeccable, and no defects are generated, with no occurrences of
rework or scrap of raw materials. The processing times described by activity, figure
15.8, are absolutely accurate and are not subject to variation of any kind.

FIGURE 15.8 REPRESENTATION OF LUCRANDO'S PRODUCTION PROCESS FOR PRODUCTS P

AND Q.

Products: The final product P is sold on the market for $ 90 per unit and has a market
limit (maximum), 100% certain, of 100 units per week. The Q product commands
higher price, $100 per unit, and has market limit (maximum), equally certain, of 50
units per week. The customer will only buy less than these quantities if the production
unit is unable to deliver them (for example, for capacity limitations).
General costs: The overheads (administrative, commercial, and other) of the
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

production unit in question that are considered fixed and known cost $6,000 per week
and cover ALL expenses and overheads, EXCEPT for raw materials and the purchased
item. For these materials, the more it is produced, the more you buy and the more you
pay for. They are therefore variable costs.

1. Is it possible to make money with this production unit?

a. If so, how much is the maximum amount of money that you can earn per
week – and with what product mix?
b. If not, how much is the least you can lose per week and with what product
mix?

15.4 SUMMARY
● OPT/TOC is a production management system that consists of at least two
fundamental elements: its "philosophy" – explained by its nine principles – and
some “proprietary” software that uses the concepts. The concepts and principles
however are very powerful in themselves.
● OPT/TOC considers that the basic goal of production systems is to “make
money” for the company they are a part of.
● The nine principles of OPT/TOC are, in themselves, intuitive and useful for
organizations wishing to achieve the main goal that OPT/TOC contemplates –
making money.
● “Making money,” according to OPT/TOC, is achieved through high levels of
performance against three operational objectives: maximizing the flow of
products sold (throughput), reducing inventory levels in the system, and
reducing operating expenses when transforming inventory into throughput. The
OPT/TOC complies with nine basic principles.
● The nine principles of the Theory of Constraints are:
1. Balance flow, not capacity.
2. The use of a non-bottleneck resource is not determined by its own
availability, but by the capacity of some system constraint (for example,
a bottleneck).
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

3. Utilizing and activating a resource are not synonymous.

4. An hour gained on a bottleneck resource is an hour gained for the whole
system.
5. An hour gained on a non-bottleneck resource is nothing, it’s just a mirage.
6. The transfer batch may not be and often should not be, equal to the
processing batch.
7. The processing batch must be variable, not fixed.
8. Bottlenecks not only determine the capacity of the entire system, but also
define their inventories.
9. Activity scheduling and production capacity should be considered
simultaneously and not sequentially. Lead times are a result of scheduling
and cannot be assumed a priori.
● The application of some of the principles of OPT/TOC can assist the decision-
making process, to some extent, without the use of OPT/TOC software.
However, some of the principles cannot be applied (mainly those related to batch
sizes and detailed scheduling) without the use of specialized software.
● It seems clear that if a company wants to take full advantage of OPT/TOC, it
should make use of some application that incorporates the principles and
concepts of OPT/TOC.
● The OPT/TOC philosophy is based on the principle that, in production systems,
there are two types of fundamentally different resources, in terms of their
importance for system performance: bottlenecks and non-bottlenecks.
● Bottlenecks, according to OPT/TOC, are the resources that constrain the
production capacity of the whole system and, therefore, should be treated in a
special way. Decisions regarding all other system resources should be subject to
decisions taken in relation to bottleneck resources.
● One of the points considered strong regarding the OPT/TOC system is its
activity scheduling system, which combines finite forward scheduling
algorithms and infinite capacity backward scheduling of non-bottleneck
resources to generate the activity schedules for the company’s various
production resources.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

● The advantages of OPT/TOC are:

1. OPT/TOC seems to have a special vocation to assist companies in
reducing their lead times and inventories. Users report reductions in lead
times of around 30% and inventories of around 40% to 75%, according to
a recent survey. OPT/TOC also seems to be a system that favors the
flexibility of the production system to change its production mix, since
mix variations can be easily evaluated by the OPT/TOC feature of
working as a simulator of the passage of orders in the factory.
2. OPT/TOC helps companies focus their attention on their problems.
Because OPT/TOC considers bottleneck resources to be worthy of special
attention and because these resources are generally few, companies are
encouraged not to disperse efforts, but to focus them on solving problems
that could compromise the performance of these bottlenecks.
3. The nine principles of OPT/TOC bring new “light” to old problems, which
contributes to a better understanding of problems and the search for new
solutions.
4. The results of the implementation of the OPT/TOC seem to come
relatively fast, because the implementation effort is less, due to the focus
of attention in few points considered critical.
5. OPT/TOC can be used as a factory simulator. In fact, OPT/TOC works
with the logic of a simulator. Questions like "what would happen if..." can
be answered more accurately with the help of a simulation tool. Also,
because it is a simulator that considers capacity constraints (at least of the
resources bottlenecks), OPT/TOC production lead times do not have to be
assumed a priori, but are actually the result of the simulation process.
6. OPT/TOC restricts the need for data with a high level of accuracy, since
only bottlenecks require highly accurate data.
The limitations of OPT/TOC are:
1. OPT/TOC is a computerized system and as such it centralizes decision-making.
There is little maneuvering area left for operators. This may not favor a greater
commitment of the workforce to the company’s objectives.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

2. OPT/TOC-based software is "proprietary" and not exactly cheap. This means

that the company that uses it will be agreeing to become somewhat dependent
on a supplier.
3. The OPT/TOC philosophy basically depends on the identification of
bottlenecks. This is not always easy to do, as many factors can contribute to
mask true bottlenecks, such as large batches, traditional production practices,
among others. If the bottleneck is wrongly identified, system performance is
compromised. There may also be a case where a factory bottleneck is "errant",
that is, it varies in resource, depending on the production mix. Although this is
not the usual case, the occurrence of wandering bottlenecks can compromise the
results of the system.
4. OPT/TOC is a system that requires a certain level of analytical skill from the
scheduler, which requires extensive training and perfect understanding of the
principles involved.
5. OPT/TOC demands that some assumptions change substantially that, for many
years, crystallized in most Western factories. This can raise resistance to its
adoption by people more resistant to change. One point that can particularly raise
controversy are the proposed new performance measures: the effort to improve
system performance in the new measures can cause performance in traditional
operational measures (such as the equipment utilization rates, for example) to be
impaired.

15.5 EXERCISES
1. What are the three basic operational objectives that OPT/TOC considers?
Discuss each of them succinctly. What are the advantages, according to
OPT/TOC advocates, of using these three objectives instead of traditional ones?
2. OPT/TOC considers that there are two types of resources within production
systems. What types are these and why would they deserve different treatment?
3. Explain the OPT/TOC principle that the production scheduling system should
balance flow and not resource capacity.
4. What is the difference between "activating" a resource and "utilizing" a
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

resource?
5. Why wouldn't the benefits be equal to reducing machine setup times for a
bottleneck machine and for a non-bottleneck machine?
6. In terms of the OPT/TOC system, what are processing and transfer batches, what
is the relationship between them for a resource and what are the main advantages
of considering both as different?
7. Briefly describe the operation of the OPT/TOC drum-buffer-rope mechanism.
8. Why is it said that OPT/TOC uses both programming logics: backward and
forward? For what situations does OPT/TOC use one and for which situations
does it use the other?
9. What is it and why is there the convenience of OPT/TOC establishing secondary
time buffers?
10.What steps should be followed to manage a production system using OPT/TOC?
Describe each one succinctly.

15.6 CLASSROOM ACTIVITIES

1. Together with your teammates, choose a situation where you or another member
is a frequent customer of an operation that suffers from long chronic queues. It
can be a popular fast-food, any public service department, the cafeteria-style
restaurant of your university or other. Identify the bottleneck resource; discuss
plausible practical ways to exploit this resource to the fullest; after that, think
about how one could “elevate” this constraint (bottleneck) of the process also
with practical and plausible suggestions.
2. Discuss with your colleagues a step-by-step process that you would use if you
were hired as consultants to implement the principles of Theory of Constraint in
a factory that is very traditional in their approach to operations management.
Think about detailed activities, including how to perform them and in what
order.

15.7 BIBLIOGRAPHY AND RECOMMENDED ADDITIONAL

READING
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner

CHASE, R.; JACOBS, R.; AQUILANO, N. J. Operations management for competitive

advantage. 10. ed. New York: McGraw-Hill, 2004.
CORRÊA, H. L. The links between uncertainty, variability of outputs and flexibility in
manufacturing systems. 1992. Tese (Doutorado) – University of Warwick,
Coventry, Inglaterra, 1992.
CORRÊA, H. L.; GIANESI, I. G. N. Just in time, MRP II and OPT: a strategic focus.
São Paulo: Atlas, 1993.
CORRÊA, H. L.; GIANESI, I. G. N.CAON, M. Planning, scheduling and production
control. 6. Ed. São Paulo: Atlas, 2018.
GOLDRATT, E.; COX, J. The goal. London: Sceduling Technology Group, 1986.
GOLDRATT, E.; COX, J. The haystack syndrome. New York: North River Press,
1990.
GOLDRATT, E.; COX, J. Computerized shop floor scheduling. International Journal
of Production Research, 26(3):443-445, 1988.
GUERREIRO, R. The goal of the company. São Paulo: Atlas, 1996.
JONES, G.; ROBERTS, M. Optimised production technology. London: IFS
Publications, 1990.
SLACK, N. D. C.; CORRÊA, H. L. The flexibilities of push and pull. International
Journal of Operations & Production Management, v. 12, no 4, 1992.
SRIKANTH, M.; UMBLE, M. Synchronous management: profit-based manufacturing
for the 21st century. Gilford: Spectrum Publishing, 1997.
VOLLMANN, T. E. et al. Manufacturing planning and control systems. 3. ed.
Chicago: Irwin, 1992.

Related websites

http://www.goldratt.com – Website of the company's "theory of constraints", Eliahu

Goldratt. Accessed: 12 Feb. 2022.

http://www.tocico.org – Theory of Constraints International Certification

Organization – organization that manages professional certification processes in
constraint theory. Accessed: 12 Feb. 2022.
Copyright Henrique Correa© - No part of this manuscript may be reproduced, posted or transmitted without the express authorization of the copyright owner