Manual Assembly Lines

• Factors favoring the use of assembly lines:

– High or medium demand for product
– Identical or similar products
– Total work content can be divided into work elements
– It is technologically impossible or economically infeasible
to automate the assembly operations
• Most consumer products are assembled on manual
assembly lines

2
Why Assembly Lines are so Productive

• Specialization of labor
– Learning curve
• Interchangeable parts
– Components made to close tolerances
• Work flow principle
– Products are brought to the workers
• Line pacing
– Workers must complete their tasks within the cycle time of
the line

3
Manual Assembly Line Defined

• A production line consisting of a sequence of workstations

where assembly tasks are performed by human workers as the
product moves along the line
• Organized to produce a single product or a limited range of
products
– Each product consists of multiple components joined
together by various assembly work elements
• Total work content - the sum of all work elements
required to assemble one product unit on the line.

4
 Manual Assembly Line

Configuration of a manual assembly line with n manually

operated workstations

5
Typical Products Made on Assembly
Lines

Automobiles Personal computers

Cooking ranges Power tools
Dishwashers Refrigerators
Dryers Telephones
Furniture Toasters
Lamps Trucks
Luggage Video DVD players
Microwave ovens Washing machines

6
Manual Assembly Line

• Products are assembled as they move along the line

– At each station a portion of the total work content is
performed on each unit
• Base parts are launched onto the beginning of the line at
regular intervals (cycle time)
– Workers add components to progressively build the
product

7
Manual Assembly Lines - Automotive
Assembly
• The assembly line dramatically
reduced the cost of manufacturing a
Model T--time to assemble a
complete chassis dropped from 12
man-hours to 1.5 man-hours
• Why? The conveyor belt controls the
speed of work, the worker spends no
time moving the parts from station to
station, The factory is carefully
designed to make sure that each
worker has all parts necessary at the
right time, any worker who is not
efficient is obvious because of the
build-up of parts at his station
8
 Assembly Workstation

• A designated location along the work-flow path at which one

or more work elements are performed by one or more
workers
• Typical operations performed at manual assembly stations

Adhesive application Electrical connections Snap fitting

Sealant application Component insertion Soldering
Arc welding Press fitting Stitching/stapling
Spot welding Riveting Threaded fasteners

9
 Work Transport Systems

• Two basic categories:

– Manual
– Mechanized
• Work units are moved between stations by the workers
without the aid of a powered conveyor
– Types:
• Work units moved in batches
• Work units moved one at a time
– Problems:
• Starving of stations
• Blocking of stations
• No pacing
10
 Manual Work Transport Systems

• For a manual assembly line, the manning level of workstation i,

symbolized Mi, is the number of workers assigned to that station;
where i = 1,2,..., n; and n = number of workstations on the line.
• The generic case is one worker: Mi = 1.
• In cases where the product is large, such as a car or a truck,
multiple workers are often assigned to one station, so that Mi > 1.
• Multiple manning conserves valuable floor space in the factory
and reduces line length and throughput time because fewer
stations are required.
• The average manning level of a manual assembly line is simply
the total number of workers on the line divided by the number of
stations; that is,
w
M= 11
n
 Manual Work Transport Systems

• where M = average manning level of the line

(workers/station),
w = number of workers on the line, and
n = number of stations on the line.
• This seemingly simple ratio is complicated by the fact that
manual assembly lines often include more workers than those
assigned to stations, so that M is not a simple average of Mi
values
– These additional workers, called utility workers, are not assigned
to specific workstations; instead they are responsible for
functions such as
(1) helping workers who fall behind,
(2) relieving workers for personal breaks, and
(3) maintenance and repair duties.
12
 Manual Work Transport Systems

• Including the utility workers in the worker count, we have

n
wu +  wi
M= i =1
n

• where wu = number of utility workers assigned to the

system; and
wi = number of workers assigned specifically to station
i for i = 1,2,..., n.

13
Mechanized Work Transport Systems

• Work units are moved by powered conveyor or other

mechanized apparatus
– Categories:
• Work units attached to conveyor
• Work units are removable from conveyor
– Problems
• Starving of stations
• Incomplete units

14
Types of Mechanized Work Transport

– Continuous transport
• Conveyor moves at constant speed
– Synchronous transport
• Work units are moved simultaneously with stop-and-go
(intermittent) motion to next stations
– Asynchronous transport
• Work units are moved independently between workstations
• Queues of work units can form in front of each station

15
Continuous Transport

Conveyor moves at constant velocity vc

16
 Synchronous Transport

All work units are moved simultaneously to their respective

next workstations with quick, discontinuous motion

17
 Asynchronous Transport

Work units move independently, not simultaneously. A work unit

departs a given station when the worker releases it. Small
queues of parts can form at each station.

18
Material Handling Equipment for
Mechanized Work Transport

Continuous transport Overhead trolley conveyor

Belt conveyor
Drag chain conveyor
Synchronous transport Walking beam
Rotary indexing machine
Asynchronous transport Power-and-free conveyor
Cart-on-track conveyor
Automated guided vehicles

19
 Line Pacing

• A manual assembly line operates at a certain cycle time

- On average, each worker must complete his/her
assigned task within this cycle time
• Pacing provides a discipline for the assembly line
workers that more or less guarantees a certain
production rate for the line
• Several levels of pacing:
1. Rigid pacing
2. Pacing with margin
3. No pacing

20
Rigid Pacing

• Each worker is allowed only a certain fixed time each

cycle to complete the assigned task
– Allowed time is set equal to the cycle time less
repositioning time
– Synchronous work transport system provides rigid pacing
• Undesirable aspects of rigid pacing:
– Incompatible with inherent human variability
– Emotionally and physically stressful to worker
– Incomplete work units if task not completed

21
Pacing with Margin

• Worker is allowed to complete the task within a specified

time range, the upper limit of which is greater than the
cycle time
• On average, the worker’s average task time must
balance with the cycle time of the line
• How to achieve pacing with margin:
– Allow queues of work units between stations
– Provide for tolerance time to be longer than cycle time
– Allow worker to move beyond station boundaries

22
No Pacing

• No time limit within which task must be completed

• Each assembly worker works at his/her own pace
• No pacing can occur when:
– Manual transport of work units is used
– Work units can be removed from the conveyor to perform
the task
– An asynchronous conveyor is used

23
 Coping with Product Variety

• Single model assembly line (SMAL)

– Every work unit is the same
• Batch model assembly line (BMAL)
– Hard product variety
– Products must be made in batches
• Mixed model assembly line (MMAL)
– Soft product variety
– Models can be assembled simultaneously without batching

24
 MMAL vs. BMAL

• Advantages of mixed model lines over batch models lines:

– No lost production time between models
– High inventories typical of batch production are avoided
– Production rates of different models can be adjusted as product
demand changes
• Difficulties with mixed model line compared to batch model
line
– Line balancing problem more complex due to differences in work
elements among models
– Scheduling the sequence of the different models is a problem
– Logistics is a problem - getting the right parts to each workstation
for the model currently there
– Cannot accommodate as wide model variations as BMAL

25
Analysis of Single Model Assembly Lines

• The assembly line must be designed to achieve a production

rate Rp sufficient to satisfy demand for the product.
• Management must decide how many shifts per week the line
will operate and how many hours per shift.
• Assuming the plant operates 50 wk/yr, then the required hourly
production rate is given by:
Da
Rp = (15.1)
50 SH
• where Rp = average production rate (units/hr),
Da = annual demand for the single product to be made
on the line (units/yr),
S = number of shifts/wk, and
H = number of hr/shift.
26
Analysis of Single Model Assembly Lines

• If the line operates 52 wk rather than 50, then Rp = Da /52SH.

• This production rate must be converted to a cycle time Tc which
is the time interval at which the line will be operated.
• The cycle time must take into account the reality that some
production time will be lost due to occasional equipment
failures, power outages, lack of a certain component needed in
assembly, quality problems, labor problems, and other reasons.
• As a consequence of these losses, the line will be up and
operating only a certain proportion of time out of the total
shift time available; this uptime proportion is referred to as the
line efficiency.
60 E
• The cycle time can be determined as: Tc = (15.2)
Rp
27
Analysis of Single Model Assembly Lines

• where Tc = cycle time of the line (min/cycle);

• Rp = required production rate, as determined from Eq. (15.1)
(units/hr); the constant 60 converts the hourly production rate
to a cycle time in minutes; and
• E = line efficiency, the proportion of shift time that the line is
up and operating.
• Typical values of E for a manual assembly line are in the
range 0.90-0.98.
• The cycle time Tc establishes the ideal cycle rate for the line:
60
Rc = (15.3)
Tc
• where Rc = cycle rate for the line (cycles/hr), and Tc is in
min/cycle
28
Analysis of Single Model Assembly Lines

• This rate Rc must be greater than the required production rate

Rp because the line efficiency E is less than 100%.
• Rp and Rc are related to E as follows:
Rp Tc
E= = (15.4)
Rc Tp

• An assembled product requires a certain total amount of time

to build, called the work content time Twc
• This is the total time of all work elements that must be
performed on the line to make one unit of the product.
• It represents the total amount of work that is accomplished on
the product by the assembly line.

29
Analysis of Single Model Assembly Lines

• It is useful to compute a theoretical minimum number of workers

that will be required on the assembly line to produce a product
with known Twc and specified production rate Rp.
• Let us make use of Eq. (15.5) in that section to determine the
number of workers on the production line:
WL (15.5)
w=
AT
• where w = number of workers on the line;
WL = workload to be accomplished in a given time period; and
AT = available time in the period.
The time period of interest will be 60 min.

30
Analysis of Single Model Assembly Lines

• The workload in that period is the hourly production rate

multiplied by the work content time of the product; that is,
WL = RpTwc (15.6)
where Rp = production rate (pc/hr), and
Twc= work content time (min/pc).
• Eq. (15.2) can be rearranged to the form Rp = 60E/Tc.
• Substituting this into Eq. (15.5), we have 60 ET
WL = wc
Tc

• The available time AT = 1 hr (60 min) multiplied by the

proportion uptime on the line; that is, AT = 60E.
• Substituting these terms for WL and AT into Eq. (15.5), the
equation reduces to the ratio Twc/Tc.
31
Analysis of Single Model Assembly Lines

• Since the number of workers must be an integer, we can

state:
T
w* = Minimum Integer  wc (15.7)
Tc

• where w* = theoretical minimum number of workers.

• If we assume one worker per station (Mi = 1 for all i, i = 1,2,...,
n; and the number of utility workers wu = 0), then this ratio also
gives the theoretical minimum number of workstations on the
line.
• Achieving this minimum value in practice is very unlikely. Eq.
(15.7) ignores several factors that exist in a real assembly line
and tend to increase the number of workers above the
theoretical minimum:
32
Analysis of Single Model Assembly Lines

– Repositioning losses.
– The line balancing problem.
– Task time variability.
– Quality problems.

33
Repositioning Losses

• Repositioning losses on a production line occur because

some time is required each cycle to reposition the worker or
the work unit or both.
• Let us define Tr as the time required each cycle to reposition
the worker or the work unit or both.
• In our subsequent analysis, we assume that Tr is the same
for all workers, although repositioning times may actually
vary among stations.
• The repositioning time Tr must be subtracted from the cycle
time Tc to obtain the available time remaining to perform the
actual assembly task at each workstation.

34
 Repositioning Losses

• Let us refer to the time to perform the assigned task at each

station as the service time.
• It is symbolized Tsi, where i is used to identify station i, i = 1, 2,..,n.
• Service times will vary among stations.
• There will be at least one station at which Tsi is maximum.
• This is sometimes referred to as the bottleneck station because
it establishes the cycle time for the entire line.
• This maximum service time must be no greater than the
difference between the cycle time Tc and the repositioning time Tr;
that is,
Max{Tsi}  Tc - Tr for i = 1,2,...,n (15.8)

35
 Repositioning Losses

• where Max{Tsi} = maximum service time among all stations

(min/cycle),
Tc = cycle time for the assembly line from Eq. (15.2)
(min/cycle), and
Tr = repositioning time (assumed the same for all stations)
(min/cycle).
• For simplicity of notation, let us use Ts to denote this maximum
allowable service time; that is,
Ts = Max{Tsi}  Tc - Tr (15.9)
• At all stations where Tsi is less than Ts workers will be idle for a portion
of the cycle, as portrayed in Figure 15.4.
• When the maximum service time does not consume the entire
available time Tc - Tr (that is, when Ts < Tc - Tr),then this means that the
line could be operated at a faster pace than Tc from Eq. (15.2).
36
 Repositioning Losses

• In this case, the cycle time Tc is usually reduced so that Tc = Ts +

Tr; this allows the production rate to be increased slightly.
• Repositioning losses reduce the amount of time that can be
devoted to productive assembly work on the line.
• These losses can be expressed in terms of an efficiency factor
as follows:
Ts Tc − Tr (15.10)
Er = =
Tc Tc
• where Er = repositioning efficiency, and the other terms are
defined above.

37
Repositioning Losses

Bottleneck Station

Tc

Tsi

Tr
1 2 3 4 n-2 n-1 n

Figure 15.4 Bottleneck station

38
The Line Balancing Problem

• The line balancing problem is concerned with assigning the

individual work elements to workstations so that all
workers have an equal amount of work.
• Two important concepts in line balancing are the separation
of the total work content into minimum rational work
elements and the precedence constraints that must be
satisfied by these elements.
• Based on these concepts we can define performance
measures for solutions to the line balancing problem

39
 Minimum Rational Work Elements

• A minimum rational work element is a small amount of work

having a specific limited objective, such as adding a component
to the base part or joining two components or performing some
other small portion of the total work content;
• A minimum rational work element cannot be subdivided any
further without loss of practicality
• The sum of the work element times is equal to the work content
time; that is ne


Twc = Tek
k =1
(15.11)

• where Tek = time to perform work element k (min), and

ne = number of work elements into which the work
content is divided; that is, k = 1,2,..., ne
40
 Minimum Rational Work Elements
• In line balancing, we make the following assumptions about
work element times:
(1) Element times are constant values, and
(2) Tek values are additive; that is, the time to perform two or
more work elements in sequence is the sum of the
individual element times
• The task time at station i, or service time as we are calling
it, Tsi, is composed of the work element times that have
been assigned to that station; that is

Tsi =  Tek (15.12)

ki

• An underlying assumption in this equation is that all Tek are

less than the maximum service time Ts.

41
 Minimum Rational Work Elements

• Different work elements require different times, and when the

elements are grouped into logical tasks and assigned to
workers, the station service times Tsi are not likely to be equal.
• Although service times vary from station-to-station, they must
add up to the work content time:
n
Twc =  Tsi (15.13)
i =1

42
 Precedence Constraints

• In addition to the variation in element times that make it

difficult to obtain equal service times for all stations, there are
restrictions on the order in which the work elements can be
performed.
• These technological requirements on the work sequence are
called precedence constraints.
• Precedence constraints can be presented graphically in the
form of a precedence diagram, Fig. 15.5, which indicates the
sequence in which the work elements must be performed.
• Work elements are symbolized by nodes, and the precedence
requirements are indicated by arrows connecting the nodes.
• The sequence proceeds from left to right.

43
Precedence Constraints

Fig.15.5 Precedence diagram for the example 15.1

44
Measures of Line Balance Efficiency

• Because of the differences in minimum rational work

element times and the precedence constraints among the
elements, it is virtually impossible to obtain a perfect line
balance.
• Measures must be defined to indicate how good a given
line balancing solution is.
• One possible measure is balance efficiency, which is the
work content time divided by the total available service time
on the line:
Twc (15.14)
Eb =
wTs

45
Measures of Line Balance Efficiency

• where Eb=balance efficiency, often expressed as a percentage;

Ts = the maximum available service time on the line
(Max{Tsi}) (mini cycle); and
w = number of workers.
• The closer the values of Twc and wTs, the less idle time on the
line.
• Eb is therefore a measure of how good the line balancing
solution is.
• A perfect line balance yields a value of Eb = 1.00.
• Typical line balancing efficiencies in industry range between
0.90 and 0.95.

46
Measures of Line Balance Efficiency

• The complement of balance efficiency is balance delay, which

indicates the amount of time lost due to imperfect balancing as
a ratio to the total time available; that is,

d=
( wT s − Twc ) (15.15)
wT s
• where d = balance delay; and the other terms have the same
meaning as before.
• A balance delay of zero indicates a perfect balance.
• Note that Eb + d = 1.

47
 Measures of Line Balance Efficiency

• Worker Requirements. We have identified three factors that

reduce the productivity of a manual assembly line.
• They can all be expressed as efficiencies:
1. line efficiency, the proportion of uptime on the line E, as defined in
Eq. (15.4)
2. repositioning efficiency, Er, as defined in Eq. (15.10)
3. balancing efficiency, Eb, as defined in Eq. (15.14)
• Together, they comprise the overall labor efficiency on the
assembly line; defined as:
• Labor efficiency on the assembly line =EErEb (15.16)
• Using this measure of labor efficiency, we can calculate a more
realistic value for the number of workers on the assembly line,
based on Eq. (15.7):
48
 Measures of Line Balance Efficiency
R pTwc Twc T
w = Minimum Integer  = = wc (15.17)
60 EEr Eb Er EbTc EbTs

• where w = number of workers required on the line,

Rp = hourly production rate (units/hr),
and Twc = work content time per product to be accomplished on
the line (min/unit).
• The trouble with this relationship is that it is difficult to
determine values for E, Er, and Eb before the line is built and
operating.
• Nevertheless, the equation provides an accurate model of the
parameters that affect the number of workers required to
accomplish a given workload on a single model assembly line.

49
 Line Balancing Problem

• The objective in line balancing is to distribute the total workload

on the assembly line as evenly as possible among the workers.
• This objective can be expressed mathematically in two
alternative but equivalent forms:
w
Minimize(wT s − Twc ) or Minimize (Ts − Tsi ) (15.18)
i =1

• subject to:
(1) T
ki
ek  Ts
• and
(2) all precedence requirements are obeyed

50
 Line Balancing Problem

• The algorithms are:

(1) Largest Candidate rule,
(2) Kilbridge and Wester method, and
(3) Ranked Positional Weights method.
• These methods are heuristic, meaning they are based on
common sense and experimentation rather than on
mathematical optimization.
• In each of the algorithms, we assume that the manning level
is one, so when we identify station i, we are also identifying
the worker at station i.

51
 Largest Candidate Rule

• In this method, work elements are arranged in descending

order according to their Tek values, as in Table 15.5.
• Given this list, the algorithm consists of the following steps:
(1) assign elements to the worker at the first workstation by
starting at the top of the list and selecting the first element that
satisfies precedence requirements and does not cause the
total sum of Tek at that station to exceed the allowable Ts; when
an element is selected for assignment to the station, start back
at the top of the list for subsequent assignments;
(2) when no more elements can be assigned without exceeding Ts,
then proceed to the next station;
(3) repeat steps 1 and 2 for the other stations in turn until all
elements have been assigned.

52
Largest Candidate Rule

Work Element Tek {min} Preceded By

3 0.7 1
8 0.6 3,4
11 0.5 9,10
2 0.4 -
10 0.38 5,8
7 0.32 3
5 0.3 2
9 0.27 6,7,8
1 0.2 -
12 0.12 11
6 0.11 3
4 0.1 1,2

TABLE 15.5 Work Elements Arranged According to Tek Value for the Largest Candidate Rule
53
Largest Candidate Rule
Station Work Element Tek {min} Station Time {min}
1 2 0.4
5 0.3
1 0.2
4 0.1 1.0
2 3 0.7
6 0.11 0.81
3 8 0.6
10 0.38 0.98
4 7 0.32
9 0.27 0.59
5 11 0.5
12 0.12 0.62

TABLE 15.6 Work Elements Assigned to Stations According to the Largest

Candidate Rule
54
Largest Candidate Rule

Balance delay d = 0.20. The line

balancing solution is presented in
Figure 15.6.

4.0
Eb = = 0.80
5(1.0)

Fig.15.6 Solution for Example 15.2, which indicates: (a) assignment of

elements according to the largest candidate rule and (b) physical
sequence of stations with assigned work elements
55
 Kilbridge and Wester Method

• It is a heuristic procedure that selects work elements for

assignment to stations according to their position in the
precedence diagram.
• This overcomes one of the difficulties with the largest candidate
rule in which an element may be selected because of a high Te
value but irrespective of its position in the precedence diagram.
• In general, the Kilbridge and Wester method provides a superior
line balance solution than the largest candidate rule (although
this is not the case for our example problem).
• In the Kilbridge and Wester method, work elements in the
precedence diagram are arranged into columns, as shown in
Figure 15.7.
• The elements can then be organized into a list according to their
columns, with the elements in the first column listed first.
56
Kilbridge and Wester Method

Columns
I II III IV V VI

0.3
5

Fig. 15.7 Work elements arranged into columns for the Kilbridge and Wester
Method
 Kilbridge and Wester Method

• If a given element can be located in more than one column,

then list all of the columns for that element, as we have done in
the case of element 5.
• In our list, we have added the feature that elements in a given
column are presented in the order of their Tek value; that is, we
have applied the largest candidate rule within each column.
• This is helpful when assigning elements to stations, because it
ensures that the larger elements are selected first, thus
increasing our chances of making the sum of Tek in each station
closer to the allowable Ts limit.
• Once the list is established, the same three-step procedure is
used as before

58
Kilbridge and Wester Method
Work element Column Tek (min) Preceded by
2 I 0.4 -
1 I 0.2 -
3 II 0.7 1
5 II, III 0.3 2
4 II 0.1 1,2
8 III 0.6 3,4
7 III 0.32 3
6 III 0.11 3
10 IV 0.38 5,8
9 IV 0.27 6,7,8
11 V 0.5 9,10
12 VI 0.12 11

Table 15.7 Work element listed according to Columns from

Figure 15.7 for the Kilbridge and Wester method. 59
Kilbridge and Wester Method
TABLE 15.8 Work Elements Assigned to Stations According to the
Kilbridge and Wester Method

Station Work Element Column Tek (min) Station Time

(min)
1 2 I 0.4
1 I 0.2
5 II 0.3
4 II 0.1 1.0
2 3 II 0.7
6 III 0.11 0.81
3 8 III 0.6
7 III 0.32 0.92
4 10 IV 0.38
9 IV 0.27 0.65
5 11 V 0.5
12 VI 0.12 0.62
60
Kilbridge and Wester Method

• Work elements are arranged in order of columns in Table 15.7.

• The Kilbridge and Wester solution is presented in Table 15.8.
• Five workers are again required, and the balance efficiency is
once more Eb = 0.80.
• Note that although the balance efficiency is the same as in the
largest candidate rule, the allocation of work elements to
stations is different

61
Ranked Positional Weights Method

• In this method, a ranked positional weight value (call it RPW for

short) is computed for each element.
• The RPW takes into account both the Tek value and its position
in the precedence diagram.
• Specifically, RPW k is calculated by summing Tek and all other
times for elements that follow Tek in the arrow chain of the
precedence diagram.
• Elements are compiled into a list according to their RPW value,
and the algorithm proceeds using the same three steps as
before.

62
Ranked Positional Weights Method
TABLE 15.9 List of Elements Ranked According to Their Ranked Positional
Weights (RPW)

Work Element RPW Tek (min) Preceded By

1 3.30 0.2 -
3 3.00 0.7 1
2 2.67 0.4 -
4 1.97 0.1 1,2
8 1.87 0.6 3,4
5 1.30 0.3 2
7 1.21 0.32 3
6 1.00 0.11 3
10 1.00 0.38 5,8
9 0.89 0.27 6,7,8
11 0.62 0.5 9,10
63
12 0.12 0.12 11
 Ranked Positional Weights Method

• Solution: The RPW must be calculated for each element.

• To illustrate,
RPW 11 = 0.5 + 0.12 = 0.62
RPW8 = 0.6 + 0.27 + 0.38 + 0.5 + 0.12 = 1.87
• Work elements are listed according to RPW value in Table
15.9.
• Assignment of elements to stations proceeds with the solution
presented in Table 15.10.
• Note that the largest Ts value is 0.92 min.
• This can be exploited by operating the line at this faster rate,
with the result that line balance efficiency is improved and
production rate is increased.
4.0
Eb = = 0.87
5(0.92)
64
TABLE 15.10 Work Elements Assigned to Stations According to the
Ranked Positional Weights (RPW) Method

Station Work Element Tek (min) Station Time (min)

1 1 0.2
3 0.7 0.90
2 2 0.4
4 0.1
5 0.3
6 0.11 0.91
3 8 0.6
7 0.32 0.92
4 10 0.38
9 0.27 0.65
5 11 0.5
12 0.12 0.62

65
Ranked Positional Weights Method

• The cycle time is Tc = Ts + Tr = 0.92 + 0.08 = 1.00; therefore,

Rc = 60 / 1.0 = 60 cycles/hr, and from Eq. (15.4),
Rp = 60 x 0.96 = 57.6 unit
• This is a better solution than the previous line balancing
methods provided.
• Some line balancing methods work better on some problems,
while other nods work better on other problems.

66
 Workstation Considerations

• A workstation is a position along the assembly line, where one or

more workers perform assembly tasks.
• If the manning level is one for all stations (Mi = 1.0 for i = 1,2,...,
n), then the number of stations is equal to the number of workers.
• In general, for any value of M for the line, n

• where L = length of the assembly line (m, ft), and L =  Lsi (15.19)
i =1
Lsi = length of station i (m, ft).
• In the case when all Ls; are equal,
L = nLs (15.20)
where Ls = station length (m, ft).

67
 Workstation Considerations

• Base parts are launched onto the beginning of the line at constant
time intervals equal to the cycle time Tc.
• This provides a constant feed rate of base parts, and if the base parts
remain fixed to the conveyor during their assembly, this feed rate will
be maintained throughout the line
1
• The feed rate is simply the reciprocal of the cycle time: f p = (15.21)
Tc
• where fp = feed rate on the line (products/min).
• A constant feed rate on a constant speed conveyor provides a center-
to-center distance between base parts given by
vc (15.22)
s =
p =vT
c c
fp
• where sp = center-to-center spacing between base parts (m/part,
ft/part), and
• vc = velocity of the conveyor (m/min, ft/min).
68
 Workstation Considerations

• One way to achieve pacing with margin in a continuous

transport system is to provide a tolerance time that is greater
than the cycle time.
• Tolerance time is defined as the time a work unit spends
inside the boundaries of the workstation.
• It is determined as the length of the station divided by the
conveyor velocity; that is, L (15.23)
Tt = s
vc
• where Tt = tolerance time (min/part), assuming that all station
lengths are equal.
• If stations have different lengths, identified by Lsi, then the
tolerance times will differ proportionally, since vc is constant.

69
 Workstation Considerations

• The total elapsed time a work unit spends on the assembly line
can be determined simply as the length of the line divided by
the conveyor velocity.
• It is also equal to the tolerance time multiplied by the number of
stations.
• Expressing these relationships in equation form, we have
(15.24)
n
L
ET = =  Tti
vc i =1
• where ET = elapsed time a work unit (specifically, the base part)
spends on the conveyor during its assembly (min).
• If all tolerances times are equal, then ET = nTi

70
 Mixed Model Assembly Lines

• A mixed model assembly line is a manual production line

capable of producing a variety of different product models
simultaneously and continuously (not in batches).
• Each workstation specializes in a certain set of assembly
work elements, but the stations are sufficiently flexible that
they can perform their respective tasks on different models.
• In this section, we discuss some of the technical issues
related to mixed model assembly lines, specifically:
(1) determining the number of workers and other operating
parameters,
(2) line balancing, and
(3) model launching.

71
 Batch-Model Assembly Lines

• A batch-model production line (BMAL) produces different

products in batches.
• There must be some similarity among the products, otherwise it
would make no sense to try to assemble them on the same
basic line.
• To determine the average production rate for the line based
on annual demand, the sum of the demands for all the products
to be made on the line is used:

D
j =1
aj

Rp =
50 S w H sh − Td
72
 Batch-Model Assembly Lines

• Where Daj=demand for product j, units/hr, and the summation is

carried out over the number of products to be made on the line
P; and Td = total downtime during the year for changeovers
between products.
• From this, the average cycle time: Tc =60E/Rp
• The average service time is the cycle time less the repositioning time,
Tr, which is assumed to be the same for all stations and products:
Ts= Tc – Tr
• To determine the number of workers on the line, a weighted average
of the work content times is used, where the weighting is based on the
annual demands for each product: P

D T
j =1
aj wcj

D
j =1
aj

w = MinimumInteger 
EbTs
73
 Batch-Model Assembly Lines

• Where Eb = anticipated line balance efficiency.

• If there is one worker per station, then this equation also
gives the number of stations.
• Now that the number of workers and stations has been
determined, each product j will be produced at its own
service time and associated cycle time and production rate.
Twcj
• The service time for product j can be determined by Tsj =
wEb
• Where Twcj = work content time for product j, min/unit.
• Then, the cycle time for product j is Tcj = Tsj + Tr
60 E
• R
The corresponding production rate for product j is: pj =
Tcj

74
 Batch-Model Assembly Lines

• Determining the number of workers on the line is again based on

a weighted average work content time as defined:
P

D T
j =1
aj wcj

D
j =1
aj

w = MinimumInteger 
EbTs

• If production rate for each model is known rather than annual

demands, then the number of workers can be determined with:
w = WL/AT,
• where workload is defined as the summation of Rpj Twcj for all j,
and AT is the available time per worker during the period under
consideration AT = 60EEr Eb

75
 Mixed-Model Assembly Lines

• The objective in mixed model line balancing is the same as for

single model lines: to spread the workload among stations as
evenly as possible.
• Algorithms used to solve the mixed model line balancing
problem are usually adaptations of methods developed for
single model lines.
• In single model line balancing, work element times are utilized
to balance the line.
• In mixed model assembly line balancing, total work element
times per shift or per hour are used.
• The objective function can be expressed as follows:
w

Minimize (w AT – WL) or Minimize =  ( AT − TT )

i =1
si
76
 Mixed-Model Assembly Lines

• where w = number of workers or stations (we are again

assuming the Mi = 1, so that n = w),
AT = available time in the period of interest (hour, shift) (min),
WL = workload to be accomplished during same period (min),
and TTsi = total service time at station i to perform its assigned
portion of the workload (min).
• The two statements in Eq. (15.31) are equivalent
• Workload can be calculated as before, using Eq. (15.28):
P
WL =  R pjTwcj
j =1

• To determine total service time at station i, we must first

compute the total time to perform each element in the workload.
• Let Tejk = time to perform work element k on product j.
77
 Mixed-Model Assembly Lines
P
• The total time per element is given by: TTk =  R pjTejk
j =1

• where TTk = total time within the workload that must be allocated
to element k for all products (min).
• Based on these TTk values, element assignments can be made
to each station according to one of the line balancing algorithms.
• Total service times at each station are computed: TTsi =  TTk
ki

• where TTsi = total service time at station i, which equals the sum
of the times of the elements that have been assigned to that
station (min).

78
 Mixed-Model Assembly Lines

• Measures of balance efficiency for mixed model assembly line

balancing correspond to those in single model line balancing
WL
Eb =
w(MaxTTsi )
• where Eb = balance efficiency,
WL = workload from Eq. (15.28) (min),
w = number of workers (stations), and
• Max{TTsi} = maximum value of total service time among all
stations in the solution.
• It is possible that the line balancing solution will yield a value of
Max{ TTsi} that is less than the available total time AT.

79
Model Launching in Mixed Model Lines

• Determining the time interval between successive launces is

refrerred to as the launching discipline.
• Two alternative launching disciplines are available for mixed-
model assembly lines: 1) Variable-rate launching; and 2) fixed-
rate launching.
• In a variable-rate launching, the time interval between the
launching of the current base part and the next is set equal to
the cycle time of the current unit.
• Since different models have different work content times and
thus different task time per station, their cycle times and launch
time intervals vary.
• The time interval in variable-rate launching can be
expressed as
Twcj
Tcv ( j ) =
wEr Eb
80
 Model Launching in Mixed Model Lines

• where Tcv(j) = time interval before the next launch in variable-rate

launching, min; Twcj = work content time of the product just
launched (model j)
• In fixed-rate launching, the time interval between two
consecutive launches is constant.
• The time interval in fixed-rate launching depends on the product
mix and production rates on the line.
• Given the hourly production schedule, as well as values of Er
which is the repositioning efficiency, and Eb, the launching time
interval is determined as
P
1
Rp
R
j =1
T
pj wcj

Tcf =
wEr Eb 81
Model Launching in Mixed Model Lines

• Where Tcf = time interval between launches in fixed-rate

launching, min; Rpj = production rate of the model j, units/hr; Twcj
= work content time of model j, min/unit; Rp = total production
rate of all models in the schedule, simply the sum of Rpj values;
P = the number of models produced in the scheduled period, j =
1 or 2.
• If manning level Mi = 1 for all i, then n can be used in place of w
in the equation.
• Congestion and idle time can be identified in each successive
launch as the difference between the cumulative fixed-rate
launching interval and the sum of the launching intervals for the
individual models that have been launched onto the line.

82
Model Launching in Mixed Model Lines

• This difference can be expressed mathematically as

m
Congestion Time or Idle Time =  Tcjh − mTcf
h =1

• Where Tcf = fixed-rate launching interval, min; m = launch

sequence during the period of interest; h = launch index
number for summation purposes; and Tcjh = the cycle time
associated with model j in the launch position h, min, calculated
as
Twcj
Tcjh =
wEr Eb

83
Fixed-Rate Launching for Two Models

• Idle time yields a negative value, indicating that the actual

sum of task times is less than the planned time for the current
launch m.
• It is desirable to minimize both congestion and idle time.
• Accordingly, the following procedure is proposed, in which the
model sequence is selected so that the square of the
difference between the cumulative fixed-rate launching
interval and the cumulative individual model-launching interval
is minimized for each launch.
• Expressing this procedure in equation form,
2
 m 
For each launch m, select j so as to minimize   Tcjh − mTcf 
 h =1 
• Where all terms have been defined above.
84