Professional Documents
Culture Documents
LOCATION
Outline
The Problem
Objective of Facility Layout
Basic Types of Layout
Product versus Process Layout
Cellular Layouts
Proximity
Assignment Problem
The Problem
In this lesson, we shall discuss how a plant or
workplace should be laid out.
Consider the problem of finding suitable locations for
machines, workstations, storage areas and aisles
within a plant.
How to find suitable locations for departments,
lounges and mail rooms and labs within a building
that houses a faculty.
The layout problem may also occur in other places
such as grocery stores, hospitals, airports, etc.
Cellular Layouts
Cellular Layouts
Machines
Enter
Worker 2
Worker 1
Exit
Key:
Product route
Worker route
Worker
3
2
1
A
10
3
12
11
C Raw materials
Parts
A
B
C
D
E
F
G
H
1 2 3 4
x x
x
x
x x
x
x
Machines
5 6 7 8 9 10 11 12
x
x
x
x
x x
x
x
x
x
x x
x
x
x
x
x
x
x x
Machines
Parts 1 2 4 3 5 6 7 8 9 10 11 12
A
x x x
x
x
B
x
x
x x
C
x
x
x
D
x x x
x
x
E
x x
x
F
x
x
x
G
x
x
x
x
H
x
x x
Machines
Parts 1 2 4 3 5 6 7 8 9 10 11 12
A
x x x
x
x
D
x x x
x
x
B
x
x
x x
C
x
x
x
E
x x
x
F
x
x
x
G
x
x
x
x
H
x
x x
Parts
A
D
B
C
E
F
G
H
1
x
x
Machines
2 4 8 3 5 6 7 9 10 11 12
x x x
x
x x x
x
x
x
x x
x
x
x
x x
x
x x
x
x
x
x
x
x x
Parts
A
D
F
B
C
E
G
H
1 2 4 8
x x x x
x x x x
x
x x
Machines
3 5 6 7 9 10 11 12
x
x
x
x
x x
x
x
x
x
x
x
x
x
Parts
A
D
F
B
C
E
G
H
1
x
x
x
Machines
2 4 8 10 3 5 6 7 9 11 12
x x x x
x x x x
x x
x
x
x x
x
x
x
x x
x
x
x
x
x
x
x x
Parts
A
D
F
C
G
B
E
H
1
x
x
x
Machines
2 4 8 10 3 6 9 5 7 11 12
x x x x
x x x x
x x
x x x
x x x
x
x x x x
x
x
x
x x x
10
12
11
Cell 2 6
Cell1
Cell 3
7
Raw materials
3
A
5
B
Production area
Office rooms
Storage
Dock area
Locker room
Tool room
O
U
A
U
U
I
X
O
E
U
From-To Chart
Assignment Method
Many methods can be used to solve the facility layout
problem. Here we discuss assignment method to minimize
material handling costs.
Suppose that some machines 1, 2, 3, 4 are required to be
located in A, B, C, D. The cost of locating machines to
locations are known and shown below. For example, if
Machine 2 is located to location C, the cost is 7 (say,
hundred dollars per month).
Location
Machine
1
2
3
4
A
10
6
8
9
B
7
4
6
5
C D
6 11
7 9
5 6
3 12
Assignment Method
The problem is to locate the machines to minimize total
material handling costs.
One solution can be (not necessarily and optimal solution)
to assign 1, 2, 3, 4 to respectively C, B, A, D. In such a
case total cost is 6+4+8+12=30 hundred dollars per
month.
Location
Machine
1
2
3
4
A
10
6
8
9
B
7
4
6
5
C D
6 11
7 9
5 6
3 12
Assignment Method
Notice in this solution that every machine is assigned to
one location and every location is assigned to one
machine. So, there is a single box in each row and each
column. Every solution will must this property.
If there are more locations than machines, dummy
machines must be added with the same cost for all
locations. Assignment method finds an optimal solution.
Location
Machine
1
2
3
4
A
10
6
8
9
B
7
4
6
5
C D
6 11
7 9
5 6
3 12
Assignment Method
1. Perform row reductions
Subtract minimum value in each row from all other row
values
2. Perform column reductions
Subtract minimum value in each column from all other
column values
3. Line Test
Cross out all zeros in matrix using minimum number of
horizontal & vertical lines. If number of lines equals number
of rows in matrix, optimum solution has been found, stop.
4. Matrix Modification
Subtract minimum uncrossed value from all uncrossed
values & add it to all cells where two lines intersect. Go to
Step 3.
Assignment Example
Location
Machine
1
2
3
4
A
10
6
8
9
Row reduction
4 1 0 5
2 0 3 5
3 1 0 1
6 2 0 9
B
7
4
6
5
C D
6 11
7 9
5 6
3 12
Column reduction
2 1 0 4
0 0 3 4
1 1 0 0
4 2 0 8
Line Test
2 1 0
0 0 3
1 1 0
4 2 0
4
4
0
8
Assignment Example
Assignment Example
A zero assignment is an assignment solution with exactly one
zero from each row and exactly one zero from each column.
After we decide to stop computation, we find a zero
assignment.
As long as the minimum number of lines is less than the
number of rows, its not possible to find a zero assignment. If
the minimum number of lines equals the number of rows, then
there exists a zero assignment.
If we mistakenly decided that the minimum number of lines to
cover all the zeros is 4, then we would stop, and attempted to
find a zero assignment with no success. Thus, the mistake
would be detected.
Assignment Example
See the matrix under line test on Slide 34. The minimum
uncovered number is 1. There are three types of numbers
and these numbers are modified in three different ways:
uncovered numbers: subtract minimum uncovered
number 1 from all uncovered numbers.
numbers covered by one line: do nothing
numbers covered by two lines: add minimum
uncovered number 1 to all numbers covered by two
lines.
The modified matrix is shown next.
Assignment Example
Modify matrix
Line Test
1
0
0
3
1
0
0
3
0
0
0
1
0
4
0
0
Location
Machines
1
2
3
4
4
5
0
8
AB
10
00
00
31
CD
04
45
00
08
0
0
0
1
0
4
0
0
4
5
0
8
# lines = # rows
so at optimal solution
Location
Machines A
1
10
2
6
3
8
4
9
B
7
4
6
5
C D
6 11
7 9
5 6
3 12
Assignment Example
5
3
A (2 ,1 )
L ( 5 , 5 ) R e c t ilin e a r
d is t a n c e
4
E u c lid e a n
d is ta n c e
x a 2 y b 2
3
A (2 ,1 )
E u c lid e a n
d is ta n c e
Example
Locating New Facilities
The new health-watch facility is targeted to serve five
census tracts in Erie, Pennsylvania. Coordinates for
the center of each census tract, along with the
projected populations, measured in thousands are
shown next. Customers will travel from the five
census tract centers to the new facility when they
need health care.
Example
Locating New Facilities
Census Tract
A
B
C
D
E
(ai,bi)
(2.5,4.5)
(3.0,2.5)
(5.5,4.0)
(5.0,2.0)
(8.0,5.0)
Population, wi
2
5
10
7
12
(ai,bi)
A
B
D
C
E
(2.5,4.5)
(3.0,2.5)
(5.0,2.0)
(5.5,4.0)
(8.0,5.0)
Weight, wi
Cumulative
Weight
(ai,bi)
D
B
C
A
E
(5.0,2.0)
(3.0,2.5)
(5.5,4.0)
(2.5,4.5)
(8.0,5.0)
Weight, wi
Cumulative
Weight
wi 5.5 ai 4.0 bi
i 1
Minimize Maximum
Rectilinear Distance
Minimize Maximum
Rectilinear Distance
Compute five numbers c1, c2, c3, c4, c5:
c1 min (ai bi )
1i n
c2 max (ai bi )
1i n
c3 min (ai bi )
1i n
c4 max ( ai bi )
1i n
c5 max(c2 c1 , c4 c3 )
Minimize Maximum
Rectilinear Distance
Define:
x1 (c1 c3 ) / 2, y1 (c1 c3 c5 ) / 2
x2 (c2 c4 ) / 2, y2 (c2 c4 c5 ) / 2
All points along the line connecting (x1,y1) and (x2,y2)
are optimal
(ai,bi)
ai+bi
-ai+bi
(2.5,4.5)
(3.0,2.5)
(5.0,2.0)
(5.5,4.0)
(8.0,5.0)
Min
c 1=
c3=
Max
c 2=
c4=
c5 max(c2 c1 , c4 c3 )
a
)
(
y
b
)
Euclidean distance =
i
i
Example
The Euclidean Distance Problem
The new health-watch facility is targeted to serve five
census tracts in Erie, Pennsylvania. Coordinates for
the center of each census tract, along with the
projected populations, measured in thousands are
shown next. Customers will travel from the five
census tract centers to the new facility when they
need health care.
Example
The Euclidean Distance Problem
Census Tract
A
B
C
D
E
(ai,bi)
(2.5,4.5)
(3.0,2.5)
(5.5,4.0)
(5.0,2.0)
(8.0,5.0)
Population, wi
2
5
10
7
12
Gravity Problem
Minimize Weighted Sum of the Squares of the
Euclidean Distances
The problem of minimizing weighted sum of the
Euclidean distances is difficult and solved by an
iterative procedure. The initial solution of the iterative
procedure is obtained from the solution of the gravity
problem which minimizes weighted sum of the
squares of the Euclidean distances. The solution of
the gravity problem is obtained using the following
n
n
formula:
wi bi
wi ai
y * i 1n
x* i 1n
wi
wi
i 1
i 1
(ai,bi)
Weight
wi
wiai
(2.5,4.5)
(3.0,2.5)
(5.5,4.0)
(5.0,2.0)
(8.0,5.0)
2
5
10
7
12
wibi
x*
wa
i 1
n
i i
w
i 1
y*
wb
i 1
n
i i
w
i 1
wi 5.72 ai 3.76 bi
i 1
a g
i 1
n
g
i 1
b g
i 1
n
g
i 1
4.5
2.5
2
4
5
2
5
7
10
12
0.61
1.67
3.68
30.71
4.62
2
2
x a1 y b1
5.72 a1 2 3.76 b1 2
g2
x a2
w2
2
y b2
w2
5.72 a2 2 3.76 b2 2
Step 3:
Modified
Trial location
xi
yi
ai
5.72
bi
wi
gi
aigi
bigi
2.5
3
5
5.5
8
4.5
2.5
2
4
5
2
5
7
10
12
0.61
1.67
3.68
30.71
4.62
41.29
1.51
5.00
18.41
168.93
36.99
230.84
2.72
4.17
7.36
122.86
23.12
160.23
3.76
Total
x = (230.84/41.29) = 5.59
y = (160.23/41.29) = 3.88
a g
b g
i
a g
i 1
n
g
i 1
b g
i 1
n
g
i 1
( x ai ) 2 ( y bi ) 2
wi
i 1
5.50 ai 2 3.99 bi 2
Application
1. Minimize weighted sum of the rectilinear distances (done)
Facilities used by many people e.g., computer centre,
gymnasium
2. Minimize maximum rectilinear distance (done)
Emergency facilities e.g., police, fire
3. Minimize weighted sum of the Euclidean distances (done)
Utilities, e.g., phone cable
4. Minimize maximum Euclidean distance (not in book)
Transmission towers e.g., radio towers
5. Minimize weighted sum of the squares of the Euclidean
distances (done)
Approximation for 3.
LESSON 23/24:
COMPUTERIZED LAYOUT TECHNIQUE
Outline
Computerized Layout Technique
A Layout Improvement Procedure, CRAFT
Distance Between Two Departments
Total Distance Traveled
Savings and a Sample Computation
Improvement Procedure
Exact Centroids
A Layout Construction Procedure, ALDEP
Following are
some
examples of
questions
addressed by
CRAFT:
Is this a good
layout?
If not, can it be
improved?
10 20 30 40 50 60 70 80 90 100
C
B
10 20 30 40 50 60 70 80 90 100
2
2
Centroid of A
=
Centroid of C
=
Distance (A,C)
=
10 20 30 40 50 60 70 80 90 100
D
(80,85)
C
(30,25)
10 20 30 40 50 60 70 80 90 100
CRAFT: Total
Distance Traveled
(a) Material handling trips
(given)
To
F ro m
A
B
C
D
A
3
6
7
B
2
7
7
C
7
5
3
D
4
7
3
(a)
CRAFT: Total
Distance Traveled
To
F ro m
A
B
C
D
3
6
7
To
F ro m
A
B
C
D
A
50
90
60
B
2
7
7
B
50
60
110
C
7
5
D
4
7
3
(a)
3
C
90
60
50
D
60
110
50
(b)
CRAFT: Total
Distance Traveled
To
F ro m
A
B
C
D
A
B
C
D
3
6
7
To
F ro m
A
B
C
D
A
50
90
60
To
F ro m
A
150
540
420
B
2
7
7
B
50
60
110
B
100
420
770
C
7
5
D
4
7
3
(a)
3
C
90
60
D
60
110
50
(b)
50
C
630
300
150
D
240
770
150
(c)
CRAFT: Savings
As stated before, given a layout CRAFT first finds the total
distance traveled as illustrated on the previous 3 slides.
CRAFT then attempts to improve the layout by pair-wise
interchanges.
If some interchange results some savings in the total
distance traveled, the interchange that saves the most
(total distance traveled) is selected.
While searching for the most savings, exact savings are
not computed. At the search stage, savings are computed
assuming when departments are interchanged, centroids
are interchanged too. This assumption does not give the
exact savings, but approximate savings only.
Exact centroids are computed later.
CRAFT: Savings
Savings are computed for all feasible pairwise interchanges.
Savings are not computed for the infeasible interchanges.
An interchange between two departments is feasible only if the
departments have the same area or they share a common
boundary. For the layout shown on Slide 7:
feasible pairs are {A,B}, {A,C}, {A,D}, {B,C}, {C,D}
and an infeasible pair is {B,D}
For the layout shown on Slide 7, savings are not computed for
interchanging B and D. Savings are computed for each of the 5
other pair-wise interchanges and the best one chosen.
After the departments are interchanged, every exact centroid is
found. This may require more computation if one or more
shape is composed of rectangular pieces. See Slides 21-25.
F ro m
A
B
C
D
A
50
90
60
B
50
60
110
C
90
60
50
D
60
110
50
To
F ro m
A
B
C
D
A
50
60
90
B
50
110
60
C
60
110
50
D
90
60
50
CRAFT: A Sample
Computation of
Savings
(a) Material handling trips
(given)
To
F ro m
A
B
C
D
A
3
6
7
B
2
7
7
C
7
5
3
D
4
7
3
(a)
CRAFT: A Sample
Computation of
Savings
F ro m
A
B
C
D
To
3
6
7
To
F ro m
A
B
C
D
A
50
60
90
B
2
7
7
B
50
110
60
C
7
5
D
4
7
3
(a)
3
C
60
110
50
D
90
60
50
(b)
CRAFT: A Sample
Computation of
Savings
F ro m
A
B
C
D
To
3
6
7
To
F ro m
A
B
C
D
A
B
C
D
A
50
60
90
To
F ro m
A
150
360
630
B
2
7
7
B
50
110
60
B
100
770
420
C
7
5
D
4
7
3
(a)
3
C
60
110
D
90
60
50
(b)
50
C
420
550
150
D
360
420
150
(c)
10 20 30 40 50 60 70 80 90 100
C
B
10 20 30 40 50 60 70 80 90 100
Let
A1 Area A1
A2 Area A2
x1 , y1
Centroid of A1
x2 , y2
Centroid of A1
50 60 70 80 90 100
A1
A2
10 20 30 40 50 60 70 80 90 100
Find the centroid of A
Rectangle
(1)
A1
A2
Total
Area
(2)
X-coordinate Multiply
of centroid (2) and (3)
(3)
(4)
A1 x1 A2 x2
A1 A2
Rectangle
(1)
A1
A2
Total
Area
(2)
Y-coordinate Multiply
of centroid (2) and (3)
(3)
(4)
A1 y1 A2 y2
A1 A2
50 60 70 80 90 100
A1
A2
10 20 30 40 50 60 70 80 90 100
Exact coordinate of area A is
ALDEP
Automated Layout Design Program
While CRAFT is an improvement procedure, ALDEP is
a construction procedure.
CRAFT requires an initial layout, which is improved by
CRAFT. ALDEP does not need any initial layout. ALDEP
constructs a layout when there is none.
Previously, in Lesson 13, we have discussed a
construction procedure and an improvement procedure
in the context of vehicle scheduling. The nearest
neighbor heuristic is a tour construction procedure
which may be improved by eliminating intersections.
ALDEP
Given
Size of the facility
The departments
Size of the departments
Proximity relationships (activity relationship chart) and
A sweep width (defined later)
ALDEP constructs a layout.
ALDEP
The size of the facility and the size of the departments are
expressed in terms of blocks.
The procedure will be explained with an example. Suppose
that the facility is 8 blocks (horizontal) 6 block (vertical).
The departments and the required number of blocks are:
Production area
14 blocks
Office rooms
10
Storage area
8
Dock area
8
Locker room
4
Tool room
4
ALDEP
A: absolutely necessary
E: especially important
I: important
O: ordinarily important
U: unimportant
X: undesirable Production area
Office rooms
Storage
Dock area
Locker room
Tool room
O
U
A
U
U
A
O
U
O
I
X
O
E
U
ALDEP
The proximity relationships are shown on the previous
slide.
ALDEP starts to allocate the departments from the upper
left corner of the facility. The first department is chosen at
random. By starting with a different department, ALDEP
can find a different layout for the same problem.
Lets start with dock rooms (D). On
D D
the upper left corner 8 blocks must
D D
be allocated for the dock area.
D D
The sweep width defines the width
D D
in number of blocks. Let sweep
width = 2. Then, dock area will be
allocated 2 4 = 8 blocks.
ALDEP
To find the next department to allocate, find the
department that has the highest proximity rating with the
dock area as given on Slide 30. Storage area (S) has the
highest proximity rating A with the dock area.
So, the storage area will be allocated next. The storage
area also needs 8 blocks.
There are only 2 2 = 4 blocks,
D D
remaining below dock area (D).
After allocating 4 blocks, the down D D
D D
wall is hit after which further
D D
allocation will be made on the
S S S S
adjacent 2 (=sweep width)
S S S S
columns and moving upwards.
ALDEP
See carefully that the allocation started from the upper left
corner and started to move downward with an width of 2
(=sweep width) blocks.
After the down wall is hit, the allocation continues on the
adjacent 2 (=sweep width) columns on the right side and
starts moving up.
This zig-zag pattern will continue.
D D
Next time, when the top wall will
D D
be hit, the allocation will continue
D D
on the adjacent 2 (=sweep width)
D D
columns on the right side and
S S S S
starts moving down.
S S S S
ALDEP
To find the next department to allocate, find the
department that has the highest proximity rating with
storage area as given on Slide 30.
Production area (P) has the highest proximity rating A with
the storage area.
The production area needs 14 blocks.
After allocating 8 blocks, the top
D D P P P P
wall is hit and the remaining 6
D D P P P P
blocks are allocated on the
D D P P P P
adjacent 2 (=sweep width)
D D P P
columns moving downward.
S S S S
S S S S
ALDEP
To find the next department to allocate, find the
department that has the highest proximity rating with
production area as given on Slide 30.
Tool room (T) has the highest proximity rating A with the
production area.
The tool room needs 4 blocks. So, 4 blocks are allocated.
Next, there is a tie. See from Slide D D P P P P
30 that both locker room (L) and
office room (O) have the proximity D D P P P P
D D P P P P
rating of U with the tool room.
D D P P T T
Ties are broken at random. So,
any of the locker room or the office S S S S T T
room can be allocated next.
S S S S
ALDEP
Lets choose locker room (L) room at random. Then, the
last department must be office room (O). The resulting
layout is shown below.
Note that since the ALDEP chooses the first department
at random and since the ties are broken at random,
ALDEP can give many solutions to the same problem.
Using the layout, the adjacency
D D P P P P O O
relationships and the proximity
D D P P P P O O
ratings, we can find an overall
D D P P P P O O
rating of each layout. Then, the
D D P P T T O O
layout with the highest overall
rating is selected. This will now be S S S S T T O O
S S S S L L L L
discussed.
ALDEP
After a layout is obtained, a score for the layout is
computed with the following conversion of proximity
relationships:
A = 43
= 64,
E
= 42 = 16
I = 41 = 4,
O
= 40 = 1
U = 0,
X
= -45 = -1024
If two departments are adjacent in the layout then the
weight corresponding to the rating between the two
departments is added to the score.
ALDEP
Lets compute the overall rating of the layout constructed.
To do this, we shall list every pair of adjacent
departments. For each pair, a letter rating will be obtained
from the activity relationship chart (a.k.a. rel chart) on
Slide 30 and then the score will be converted to a numeric
score using the conversion scheme on the previous slide.
Adjacent departments:
(D,S)
(D,P)
(S,P)
(S,T)
(S,L)
(P,T)
(P,O)
(T,L)
(T,O)
(L,O)
D
D
D
D
S
S
D
D
D
D
S
S
P
P
P
P
S
S
P
P
P
P
S
S
P
P
P
T
T
L
P
P
P
T
T
L
O
O
O
O
O
L
O
O
O
O
O
L
ALDEP
Adjacent
Proximity
Numeric
Departments Ratings (Slide 30) Scores (Slide 37)
(D,S)
A
43=64
(D,P)
I
41=4
(S,P)
A
43=64
(S,T)
O
40=1
(S,L)
U
0
(P,T)
A
43=64
(P,O)
O
40=64
(T,L)
U
0
(T,O)
U
0
(L,O)
X
-45= -1024
Total = -763
ALDEP
The process is repeated several times and the layout
with the highest score is chosen.
Notice the large negative weight associated with X
ratings.
If the departments which cannot be next to each other,
are adjacent in a layout, then the layout score reduces
significantly.
This is important because ALDEP also uses a cut-off
score (if not specified by the user this cut-off is zero) to
eliminate any layout which has a layout score less than
the cut-off score.