THEORY OF CONSTRAINT

1
 Learning Outcome
• After attending this course, students are
expected to:
1. Understand the concept of the theory of
constraints and the basic principles in
the theory of constraints
2. Able to explain the relationship between
operational performance measures and
financial performance measures and
company goals
3. Understand the systematic steps of
Optimize Production Technology
2
 Introduction
• Developed by Goldratt in the early 70s
• Starting from software to create a production
schedule called OPT (Optimized Production
Technology)
• Later it was developed into Theory of
Constraint (TOC), and Drum Buffer Rope
Technique
• Also known as Synchronized manufacturing: all
manufacturing processes work together to
achieve company goals

3
 Company Goal

• Goal: the goal of a firm is to make money

• Goal (?): create jobs, increase sales, increase
market share, develop technology, manufacture
high quality products
• This (?) goal does not guarantee long-term
survival company
• Goal vs. means

4
 Performance Measures
Classic Financial
Measurement

1. Net Profit New Operational

2. Return on Investment Measurement
3. Cash Flow according to TOC
1. Throughput (T)
2. Inventory (I)
3. Operating Expenses (OE)
5
 New Operational Measurements

• Throughput:
– The rate at which the entire system generates money
through sales
– All the money coming into the system
– “The number of deliveries of a product or service to
customer” (Non-profit company)
– The world is not interested in the storms you
encountered, but did you bring in the ship (William
McFee dalam Dettmer [1997]

6
 New Operational Measurements

• Inventory:
– This is money that the system has invested in
purchasing things it intends to sell;
– It measures inventory in terms of material cost
only, without accounting for labor and overhead
• Operating Expenses:
– This is the money that the system spends in
order to turn inventory into throughput,
including all labor, overhead, and other
expenses

7
 Company Goal

• Production (operational) activities must have

high throughput performance and at the same
time low inventory and operating expense
• If this can be achieved, the company will be
closer to the company's goals (make money).
• Back to the productivity concept: output per
labor hour
• Productivity: all actions that bring the
company closer to its goals

8
 Relation between Financial and
Operasional Measurement
NET PROFIT ROI CASH FLOW
increase increase increase

T I OE
increase decrease decrease

Net Profit = T - OE Cash Flow= T – I - OE

ROI = (T – OE)/I
9
 T-I-OE Priority
• From the formula, it can be seen that related to
financial performance, in order to achieve the
goal (goal), the steps taken operationally
(production) are:
• First, increase Throughput
• Second, reduce Inventory
• Third, reduce operating expenses

10
 Concept of Constraints (1)
• The problem faced to increase T and suppress I
and OE is the existence of constraints in the
system
• This limitation becomes an obstacle to these
efforts
• It must be recognized and then addressed so
that efforts to increase T and suppress I and
OE can be carried out
• Also called constraint management: the
constraint determines Throughput

11
 Concept of Constraints (2)
• There are 3 (three) types of constraints
• Material Constraints: limitations due to lack of
material, rarely happens
• Resource Constraints: limitations due to lack of
capacity in machines, people, skills and
markets. More common than material
constraints
• Policy Constraints: limitations due to settings
or conditions that affect organizational
behavior. For example, the batching rules, how
to set-up, etc. This limitation is the most
common (90%) 12
 Definition
• Capacity: time available to carry out production
activities
• Balanced vs. unbalanced capacities:
• Capacity should not be balanced, but flow
should be balanced
• If the flow is balanced then the capacity will
not be balanced
• Bottleneck resource (BR): any resource
whose capacity is equal to or less than the
demand placed upon it
• Non-bottleneck resource (NBR): any resource whose
capacity is greater than the demand placed upon it 13
 Definition(2)
• Capacity-constrained resource (CCR): a resource that has
become a bottleneck as a result of inefficient utilization.
• Setup time: the time the part spends because the
resources that will process the part are being set up
• Process time: the time the part is being processed
• Queue time: part waiting time (queue) because the
machine is busy working on other parts or doing other
things

14
 Definition (3)

• Wait time: waktu tunggu part karena part lain

yang akan dirakit bersama belum ada
• Idle time: waktu tidak terpakai (unused time),
yaitu, waktu siklus dikurangi nilai penjumlahan
waktu setup, waktu proses, waktu antri dan
waktu tunggu

Wait Queue Setup Processing Idle

Work center Resource time

15
 Definition (4)
Work center Resource time

W Q S P

a. Bottleneck resource (BR)

W Q S P I

b. CCR

W Q S P I

c. Non-bottleneck resource (NBR)

16
Basic Manufacturing building block(1)

a. BR feeding NBR d. B and NB have independent markets

Market Market

b. NBR feeding BR Market

Market

c. Output B and NB assembled

BR
Final
assembly Market
NBR

17
 Basic Manufacturing building block(2)

Y X Y Y Output

15 minutes 20 minutes 10 minutes 12 minutes

4 unit/hour 3 unit/hour 6 unit/hour 5 unit/hour

How many units are produced in an hour? 3

18
Basic Manufacturing building block(3)

X Y Market
300 units 300 units
(200 hours) (150 hours)
X: 200/200= 100%
Y: 150/200= 75%

WIP
Y X Market

Y works 75% or WIP (work in process) occurs

19
Basic Manufacturing building block(4)

X
Market
Y Spare parts

Y works 75% or waiting spare parts

X Market

Y Market
Finished
good
Y works 75% or accumulated finished good
20
 Number of bottleneck resource

Is it possible to have more than 1 bottleneck resource

in a system?

Y Y X Y X M

15 20 10 12 8 (1)
unit/hour unit/hour unit/hour unit/hour unit/hour
15 unit/hour 20 unit/hour 10 unit/hour 12 unit/hour 10 unit/hour (2)

Condition 1: How many number of 2 machine

bottleneck resources?
Condition 2: How many number of 1 machine
bottleneck resources? 21
 OPT Rules (1)

• Rule 1: the level of utilization of a non-

bottleneck is determined not by its own
potential, but by some other constraints in the
system
– Non-bottleneck resources (NBR) should not be utilized
to 100% of their capacity
– They should be scheduled and operated based on
other constraints in the system
– If it is done, NBR would not produce more that the
bottlenecks can absorb, thereby preventing an
increase in inventory and operating expenses

22
 OPT Rules (2)
• Rule 2: utilization and activation of a resource
are not synonymous
– Utilization is concerned with efficiency
– Activation is concerned with effectiveness
– If a resource utilization is 100% but the BR can only
absorb 60% then there will 40% inventory
– In this situation, the resource has 100% efficiency,
but the system-wide effectiveness is only 60%

23
 OPT Rules (3)

• Rule 3: an hour lost at a bottleneck is an hour

lost for the total system
• Rule 4: an hour saved at a non-bottleneck is just
a mirage
– Maximizing (100%) utilization of bottleneck
resource
– Reduce the amount of set-ups, i.e. make the
batch sizes as large as possible, in order to
maximize the amount of productive time a
bottleneck

24
 OPT Rules (4)

• Rule 5: bottlenecks govern both throughput and

inventory in the system
– Inventory (khususnya Work In Process/WIP)
merupakan fungsi dari jumlah persediaan yang
diperlukan untuk membuat bottleneck resources sibuk
(utilisasi 100%)

25
 OPT Rules (5)
• Rule 6: the transfer batch may not, and many
times should not, be equal to the process batch
• Rule 7: the process batch should be variable, not
fixed
– It implies that the process batch size at different work centers should
not be the same
– Process batches are a function of the schedule and potentially vary by
operation and over time
– Lot size is established dynamically for each operation and balances
inventory cost, set-up costs, component flow requirements, and the
needs for managerial control and flexibility.
– Some operations might be bottlenecks and may require large process
batches
– Non-bottlenecks may require small process batches in order to reduce
lead time and the resulting inventory.
26
27
 OPT Rules (6)

• Rule 8: capacity and priority should be

considered simultaneously, not sequentially
• Rule 9: balance flow not capacity
• Rule 10: the sum of local optima is not equal to
the optimum of the whole

28
 Systematic Steps of OPT

1. Identify the system’s constraints

2. Decide how to exploit the system’s constraint
3. Subordinate everything else to step 2
4. Elevate the system’s constraint
5. If a constraint is broken, go to step 1

29
 Example (1)

Consider the production process described in the

figure below. Two products, P and Q are
manufactured; weekly demand is 100 units for P
and 50 units for Q. The selling price for P and Q,
respectively, is $90 and $100. There are four
workcenters: A, B, C and D. Each center has one
machine that can operate up to 2400 minutes per
week. Three types of raw materials are required.
Raw materials costs and the routings and
processing times at each workcenter are shown
below.
30
Example (1)

31
 Example (3)
Step 1. Identify the system’s constraints

Kapasitas yang diperlukan Kapasitas Persentase

Resource
P Q Total Tersedia Beban
A 1500 500 2000 2400 83%
B 1500 1500 3000 2400 125%
C 1500 250 1750 2400 73%
D 1000 250 1250 2400 52%

The obvious bottleneck is workcenter B.

In order to satisfy the market demand for P and Q, it
should have an additional 25% capacity.

32
 Example (4)
Step 2. Decide how to exploit the system’s constraint
Product P Q
Price 90 100
Raw material cost 45 40
Contribution 45 60
Time spent in WC. B 15 30
Contribution to time spent ratio 3 2

TOC focused on the maximizing the utilization of the constraint

resource to achieve the goal (make money). Exploiting the
workcenter B means maximizing the return of its utilization in any
single minute. Therefore, it is more advantageous to produce product
P then Q. So, P will be produced in 100 units (requires 1500 minutes)
and the remaining time is used to produce Q = 30 units in 900
minutes.

33
 Example (5)
Step 3. Subordinate everything else to step

DBR (drum-buffer-rope) technique

Step 4. Elevate the system’s constraint: setup

reduction, preventive maintenance etc.
If the demand of product P increases
from 100 to 150 units then product Q
should be manufactured in another
resource.

34
 Example (6)
Step 5. If a constraint is broken, go to step 1
If demand of P and Q increase to 32 units and 66
units, and after some improvement the processing
time in workcenter B cuts down to one third of the
previous processing time (5 minutes per unit), then

The workcenter A is the

Resource Beban
constraint resource
A 110
B 55
C 96,25
D 68,75
35
 Linear Programming Model(1)

Max Z = 45P + 60Q S.t.

15 P + 10 Q  2400 (WC A)
15 P + 30 Q  2400 (WC B)
15 P + 5 Q  2400 (WC C)
10 P + 5 Q  2400 (WC D)
P 100 (demand of product P)
Q  50 (demand of product Q)
P, Q  0 (non-negative constraints)

36
OPT Framework

37
 OPT Software (1)

• Buildnet Module: An MRP II data base is used to construct a consolidated

network for each end product. This data base includes bill of material,
routings, inventories, workcenters, market requirements and other data.
Constructing the end-product data network is the most time-consuming
element of OPT
• Serve Module: The serve module calculates a load profile and average
utilization for each resource
• Split Module: This module divides the network into two areas, critical and
noncritical resources. It also allocates time buffer at the appropriate
places.
• OPT Module: OPT is a finite scheduler (i.e., bottlenecks are scheduled first
that “beats the drum” of the bottleneck. By using a good heuristic, a
realistic (i.e., one that considers the constraints) master production
schedule is generated. This includes a work schedule report, a load profile
for each bottleneck (or CCR), and a forecast of order completion

38
OPT Software (2)

39
 OPT Illustration (1)
Gizmo-Stools Inc. manufactures two types of stool,
namely, a four-legged stool and a three-legged stool.
The product structures are shown below
F-456 F-449
Stool A Stool B

P-455 A-454 P-455 A-448

Leg Seat A Leg Seat B
(4) (1) (4) (1)

P-453 P-452 P-447 P-452

Frame A Cushion Frame B Cushion
(1) (1) (1) (1)

R-450 R-450 R-451 R-450

Raw frame Cushion Raw frame Cushion
(1) material (1) material 40
(1) (1)
 OPT Illustration (2)
Process requirements for Stool B
Number of operations 3 (All times in hours)
Operation number 10 20 30
Description Assemble legs Paint Stool Inspect Stool
to stool
Set-up time 0.5 0.75 0.5
Processing time 0.25 0.35 0.2
Operator time 0.25 0.35 0.2
Transport time 1.00 1.00 1.00
Work center Assembly Painting Inspect
Next operation 20 30 Stock room

If the demand for both Stools, A and B are 50 each and should be delivered in
hour 40. 41
 OPT Illustration (3)
• Stool product network

Customer orders
and/or forecast Finished stool

Inspect finished stool

Parts and
processing Paint Station:
sequences Paint assembled stool Bottleneck
(longest SU+RT)

Assemble legs and seat

Raw materials Seats Legs 42

 OPT Illustration (4)

▪ Assume Stool B is more advantageous than Stool

A, Stool B will be produced first
▪ Transfer batch = 1; using backward scheduling
one obtains:

43
 OPT Illustration (5)
▪ It is evident that painting station is the
bottleneck which causes stack of WIP
▪ Painting and inspection will be critical networks.
Forward scheduling is performed, one obtain:

44
 OPT Illustration (6)

• Scheduling is performed for Stool B first, then it

is followed by Stool A
• The system performance can be better is mixed
model production is applied
• Stool B and Stool A are produced alternately
• It requires fast setup to change from producing
A to producing B and vice versa

45
 DBR Technique (1)

▪ OPT is implemented in a software

▪ For easy application – without a
software – it is developed a scheduling
technique that is called Drum-Buffer-
Rope (DBR)

46
 DBR Technique (2)

• Drum: the constraint resource that control

the pace of the system
• Buffer: it is used to protect the bottleneck
from fluctuations and variations in the
feeding rate to the bottleneck
• Rope: a feedback loop which connected the
bottleneck to the raw material dispatching
point

47
DBR Technique (3)

48
DBR with two feedback loops
 DBR Technique (4)

• Steps of DBR scheduling:

• Scheduling the bottleneck (constraint)
• Operations are then scheduled backward from
the bottleneck to the raw material dispatching
point and
• Then forward to shipping (the forward schedule
provides estimates for customers’ deliveries
• DBR can be regarded as a combination of both push
and pull.
• The dispatching of materials is governed by pull
and form there on it is pushed
50
 Buffer (1)

▪ Stock vs Time Buffers

▪ Time buffers: designed to protect the throughput
of the system from the internal disruptions that
continually occur in manufacturing environment
▪ Stock buffers: designed to improve the
responsiveness of the operation to market
demand. This is accomplished by holding
inventories of finished products in anticipation of
future market demand. This allows for orders to be
filled in less than the normal production lead time.
▪ Buffer size: ¼ to ½ of the Manufacturing Lead Time
(MLT)
51
Buffer (2)

52
Buffer (3)

53