GROUP TECHNOLOGY

PART I
Andrew Kusiak Intelligent Systems Laboratory 2139 Seamans Center The University of Iowa Iowa City, Iowa 52242 - 1527 Tel: 319 - 335 5934 Fax: 319 - 335 5669 andrew-kusiak@uiowa.edu http://www.icaen.uiowa.edu/~ankusiak
The University of Iowa Intelligent Systems Laboratory

Contents
INTRODUCTION CLUSTER ANALYSIS METHOD BRANCHING ALGORITHMS ASSIGNMENT OF PARTS TO THE EXISTING MACHINE CELLS SUMMARY
The University of Iowa Intelligent Systems Laboratory

Grouping Parts

What is Group Technology?

Simplification Approach Decomposition Approach

Parts Before Grouping

Grouped Parts
PF - 1 PF - 2

PF - 3

PF - 4

Both accomplished through: Grouping Parts Grouping Routes

The University of Iowa Intelligent Systems Laboratory The University of Iowa Intelligent Systems Laboratory

Grouping Routes
Ungrouped Parts and Machines
P1 PF - 1 M1 P3 M4 P1 P2 M6 M2 M5 P3 P2 M3 P5 M6 P4 P5 M1

Group Technology
Reduced production lead time
MC - 1 M2

Grouped Parts and Machines

P4

PF - 2 M3 M4

MC - 2

M5

Parts

Machines

Parts Families

Machines Cells

Reduced work-in-process Reduced labor Reduced tooling Reduced rework and scrap materials Reduced setup time Reduced order time delivery Improved human relations Reduced paper work
Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Page 1
1

Visual Method
Two Basic Methods of Group Technology:
PF - 1 PF - 2

Classification
- Visual method - Coding method
PF - 3 PF - 4

Cluster analysis
The University of Iowa Intelligent Systems Laboratory

Parts before grouping

The University of Iowa

Grouped parts
Intelligent Systems Laboratory

Monocode system and parts monocode Coding Method

A code is assigned to each part based on its characteristics, e.g., geometry, material, etc.
Digit 1 Monocode system 1. Steel 2. Copper 3. Bronze . . .

Sample monocode 3 1 2

Digit 2

Round and straight

1 inch thick and straight Single through center hole

Three basic types of coding systems:

Monocode Polycode Hybrid
The University of Iowa Intelligent Systems Laboratory Digit 3

No center hole

Single through center hole

1 Two center holes

Round and straight Bronze

The University of Iowa

Intelligent Systems Laboratory

?
Question:
Is the classification and coding method exact or ambiguous?

CLUSTER ANALYSIS APPROACH

Cluster analysis is concerned with grouping of objects into homogeneous clusters (groups) based on the object features. The application of cluster analysis in manufacturing leads to grouping parts into part families (PF) and machines into machine cells (MC). The result of this grouping is implemented as: - Physical machine layout, or - Logical (virtual) machine layout
The University of Iowa Intelligent Systems Laboratory

Ambiguous!
The University of Iowa Intelligent Systems Laboratory

Page 2
2

Questions:

Implementation of GT:
Physical Machine Layout
Parts P1 Machines MC - 1 PF - 1 M1 P3 M4 P1 P2 M6 M2 M5 P3 P2 M3 P5 Ungrouped parts and machines Grouped parts and machines
Intelligent Systems Laboratory

What is the main difference between the classification and coding approach and the cluster analysis approach? Which of the two GT approaches requires information at an earlier phase of the product development cycle?
The University of Iowa Intelligent Systems Laboratory

M1

M2

P4

PF - 2 M3 P5 M4

MC - 2

M6

P4

M5

The University of Iowa

Implementation of GT:
Logical Machine Layout
Parts P1 PF - 1 M1 P3 M4 P1 P2 M2 P4 M2 M5 P5 P2 M3 P5 M6 P3 P2 M3 M6 PF - 2 M5 M1 M4 Machines MC - 1 MC - 2

Formulations of the GT Problem

Matrix formulation Mathematical programming formulation Graph formulation

Ungrouped parts and machines

The University of Iowa

Grouped parts and machines

Intelligent Systems Laboratory The University of Iowa Intelligent Systems Laboratory

Mutually separable clusters PF -1

PF-2 2 4 5

Matrix Formulation
1 Part number 1 1 Machine 2 number 3 4 1 1 1 1 2 1 1 1
MC-2 3 1 1 PF -1 PF-2 2 4 5

2 1

4 1

5 1 2 MC-1 4

1 1 1

3 1 1

2 3 4

1 1 1

1 1 MC-2 1

4 1

5 1
2 MC-1 4 1

1 1 1

3 1 1

1 3

1 1

1 1

Partially separable clusters

1 1 MC-1 { 2 3 MC-2 { 4 1 1
Intelligent Systems Laboratory

2 1 1

5 1

1 1

Bottleneck part

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Page 3
3

Partially separable clusters

1 1 MC-1 { 2 3 MC-2 { 4 1 1 1 1 1 1 1 1 2 1 3 4 5 1

Partially separable clusters

Part number 1 2 1 1 1 3 4 5 6

Bottleneck part 5

1 2 Machine 3 number 4 5

1 1 1

1 1 1

1 1 1

1 1 1

bottleneck machine 3
1 MC-1 { 2 3(1) 3(2) MC-2 { 4 5

PF-1 1 1 1 1 2 1 1 1 1 1 1 1 3 4

PF-2 5 6

The bottleneck part 5 prevents the decomposition of a machine-part incidence matrix into disjoint submatrices
The University of Iowa Intelligent Systems Laboratory

1 1 1

1 1 1

The University of Iowa

Intelligent Systems Laboratory

Question:

?
1 2 1 1 1 1 1 1 1 3 4 5 1 1 1

Three actions can be taken to deal with the bottleneck part 5

1. It can be machined in one machine cell and transferred to the other machine cell by a material handing carrier. 2. It can be machined in a functional facility. 3. It can be subcontracted.
The University of Iowa Intelligent Systems Laboratory

What can be done to deal with the bottleneck part 5?

1 MC-1 { 2 3 MC-2 { 4

The University of Iowa

Intelligent Systems Laboratory

Solving the GT problem represented with a binary matrix: Similarity coefficient methods Sorting based algorithms Bond energy algorithm Cost-based method Cluster identification algorithm Extended cluster identification algorithm
The University of Iowa Intelligent Systems Laboratory

Cluster Identification Algorithm:

Special Case of the GT Problem

The University of Iowa

Intelligent Systems Laboratory

Page 4
4

Example
Incidence matrix
1 2 3 *

Part
1 2 3 4 5 6 7 8 * * * * * * * * * * * * * *

Machine 4 *
5 6 7
The University of Iowa

1 2 3 4 5 6 7

1 2 3 4 5 6 7 8 * * * * * * * * * * * * * * * *
Intelligent Systems Laboratory

Intelligent Systems Laboratory

The University of Iowa

Next Step
1

1 2 3 4 5 6 7 8 * * * * * * * * * * * * * * * *

1 4 6 7 2 3 4 6 * * * * * * *

Delete all double-crossed elements

2 3 4 5 6 7

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

Resultant Matrix
Iteration 2

1 4 6 7 2 3 4 6 * * * * v1 v6
4 7 * * * v4 v7

* * *

h2 h4

1 4 6 7 2 3 4 6 * * * * *
Iteration 3

* *

3 6

h3 h6

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

Page 5
5

Final decomposition result

CI Algorithm Repeated
(For notes follow up)

CC-1 CC-2 CC-3 2 3 5 8 1 6 4 7 RC-1 1 * * * 5 * * * 7 * * * * RC-2 2 * * 4 * * * * RC-3 3 6 *

The University of Iowa Intelligent Systems Laboratory The University of Iowa Intelligent Systems Laboratory

Step 0. Set iteration number k = 1. Step 1. Row 1 of matrix (6) is selected and horizontal line h1 is drawn. Step 2. Three vertical lines v2, v3, and v5 are drawn for each single-crossed entry 1.
1 1 2 3 4 5 6 7 1 1 1 1 1 1 v
2

Step 3. Two horizontal lines h5 and h7 are drawn through the single-crossed entries of last matrix
1 1 2 3 4 5 6 7 1 1 1 1 1 1 v
1
2

2 1

3 1

5 1

6 1

8 h
1

1 1 1 1 1 1
5

2 1

3 1

5 1

6 1

8 h
1

1 v
3 1 1
3

v
4 5 1

1 1 1 1 1 1
5

2 1

6 1

8 h1

1 v
3

1 2 3 4 5 6 7

1 1 1 1 1 v2 1 v3 1 v5 1

1 1 1 1 h5 h7

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

Step 4. Since the entries (5, 8) and (7, 8) are single-crossed, the vertical line v8 is drawn.
1 1 2 3 4 5 6 7 1 1 1 1 1 1 v
2

Step 5. Removing all double-crossed entries.

1 2 3 4 6 1 1 1 1 1 4 6 1 1 7

2 1

3 1

5 1

6 1

8 h
1

1 1 1 1 v
5

1 1 v
8

h h

5 7

1 v
3

In the second iteration (k = 2), Steps 1 through 4 are performed for matrix above. This iteration results in the following incidence matrix:
1 4 1 1 1 v
1

6 1

7 h 1 h
2 4

As there are no more single-crossed entries, all double-crossed entries 1 form: Machine cell MC-1 = {1, 5, 7}, and Part family PF-1 = {2, 3, 5, 8}.
The University of Iowa Intelligent Systems Laboratory

2 3 4 6

1 v

6 Intelligent Systems Laboratory

The University of Iowa

Page 6
6

In the third iteration (k = 3), removing double-crossed entries from matrix below
4 3 6 1 1 v
4

7 1 v h h3
6 7

Conceptual design of a cellular manufacturing system

Manufacturing system Machine cell MC - 1 Machines: 1, 5, 7 Machine cell MC - 2 Machines: 2, 4 Parts: 2, 3, 5, 8 Machine cell MC - 3 Machines: 3, 6 Parts: 4, 7

From this matrix MC-3 = {3, 6} and PF-3 = {4, 7} are obtained. The final clustering result with three machine cells and part PF-1 PF-2 PF-3 families:
2 1 MC-1 { 5 7 2 MC-2 { 4 3 MC-3 { 6 The University of Iowa 1 1 3 1 1 1 5 1 1 1 8 1 1 1 1 1 1 1 1 1 1 6 4 7

Parts: 1, 6

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

Extended CI Algorithm
Step 0. Set iteration number k = 1. (k) Step 1. Select those machines (rows of matrix [aij] ) that based on the user's expertise, are potential candidates for inclusion in machine cell MC-k. Draw a horizontal line hi through each row of matrix [aij](k) corresponding to these machines. In the absence of the user's expertise any machine can be selected. (k) Step 2. For each column in [aij] corresponding to entry 1, single-crossed by any of the horizontal lines hi , draw a vertical line vj . (k) Step 3. For each row in [aij] corresponding to the entry 1, single-crossed by the vertical line vj, drawn in Step 2, draw a horizontal line hi . Based on the machines corresponding to all the horizontal lines drawn in Step 1 and Step 3, a temporary machine cell MC-k , is formed.
The University of Iowa Intelligent Systems Laboratory

If the user's expertise indicates that some of the machines cannot be included in the temporary machine cell MC-k, (k) erase the corresponding horizontal lines in the matrix [aij]. Removal of these horizontal lines results in machine cell MC'-k . Delete from matrix [aij] (k )parts (columns) that are to be manufactured on at least one of the machines already included in MC-k. Place these parts on the list of parts to be manufactured in a functional machining facility. Draw a vertical line vj through each single-crossed entry 1 in [aij](k) which does not involve any other machines than those included in MC-k .
(k)

Step 4. For all the double-crossed entries 1 in [aij] , form a machine cell MC-k and a part family PF-k
The University of Iowa Intelligent Systems Laboratory

Step 5. Transform the incidence matrix [aij] into [aij] (k+1) by removing all the rows and columns included in MC-k and PF-k, respectively. Step 6. If matrix [aij] (k +1)= 0 (where 0 denotes a matrix with all elements equal to zero), stop; otherwise set k = k + 1 and go to step 1.

(k)

Example: Extended CI Algorithm

Constraints: Max |MC| = 4 Machines 1 and 4 in the same cell Part number
1 1 2 3 4 5 6 7 1 1 1 1
The University of Iowa

2 1

3 1

5 1

7 1

10

11 1 Machine number

1 1 1 1 1 1 1 1 1 1 1 1 1

The University of Iowa

Intelligent Systems Laboratory

Intelligent Systems Laboratory

Page 7
7

Working Incidence Matrix 1

Part number 1 1 2 3 4 5 6 7 1 1 1 1 1 1 1 1 1 1 2 1 3 1 1 1 1 1 1 1 1 1 4 5 6 7 1 1 Machine number 8 9 10 11

Step 0. Set iteration number k = 1. Step 1. Since user's expertise indicates that machines 1 and 4 should be included on machine cell MC-1, two horizontal lines h1 and h 4 are drawn.
Part number 1 1 2 3 4 5 6 7 1 1 1 1
The University of Iowa

2 1

3 1

5 1

7 1

10

11 h1 1

1 1 1 1 1 1 1 1 1 1 1 1 1 h4

The University of Iowa

Intelligent Systems Laboratory

Intelligent Systems Laboratory

Step 2. For columns 1, 2, 3, 6, and 7 crossed by the horizontal lines h 1 and h 4, five vertical lines, v1, v2, v3, v6, and v7 are drawn.
Part number 1 1 2 3 4 5 6 7 1 1 1 1 v1 v2 v3 1 1 1 1 1 v6 1 v7 2 1 3 1 1 1 1 1 1 1 1 1 h4 4 5 6 7 1 1 8 9 10 11 h1

Step 3. Three horizontal lines, h2, h6, and h7 are drawn through rows 2, 5, and 7 corresponding of the single-crossed elements 1.
1 1 2 3 4 5 6 7 1 1 1 1 v1 v2 v3 1 1 1 1 1 v6 1 v7
Intelligent Systems Laboratory

2 1

3 1

Part number 5 6 7 1 1

10

11 1 h1 h2 h4

1 1 1 1 1 1 1 h6 h7

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Working Incidence Matrix 2

Part number 1 1 2 3 4 5 6 7 1 1 1 1 1 1 1 1 1 1 2 1 3 1 1 1 1 1 1 1 1 1 4 5 6 7 1 1 Machine number 8 9 10 11

A temporary machine cell MC'-1 with machines {1, 2, 4, 6, 7} is formed. Machines 2 and 6 are excluded from in MC'-1 as they include less double-crossed (committed) elements than machine 7. The horizontal lines h2 and h6 are erased.

2 1 2 3 4 5 6 7 1

3 1

5 1

7 1

10 11 h1 1 1 h4

1 1 1 v2 v3

1 1 1 1 v6 1 v7 1

h7
Intelligent Systems Laboratory

M = 1, 4, 7
The University of Iowa Intelligent Systems Laboratory

The University of Iowa

Page 8
8

Step 4. The double-crossed entries 1 of matrix indicate: Machine cell MC-1 = {1, 4, 7} and Part family PF-1 = {2, 3, 6, 7} Step 5. Matrix is transformed into the matrix below.

Step 6. Set k = k + 1 = 2 and go to Step 1. The second iteration (k = 2) results in: Machine cell MC-2 = {2, 3, 5, 6} and Part family PF-2 = {5, 8, 10, 11}

5 2 3 5 6 1

10 1

11 1 h2 h3 h5 h6

5 2 3 5 6
The University of Iowa

10 1

11 1

1 1 1 1 1

1 v5

1 1 v8

1 v10 v11

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

The final result

PF - 1 2 MC - 1 MC - 2 1 4 7 2 3 5 6 1 3 1 1 1 6 1 1 7 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 5 8 PF - 2 10 11 1 4 9

Conceptual layout of the manufacturing system

Manufacturing system Machine cell MC - 1 Machines: 1, 4, 7 Parts: 2, 3, 6, 7 Machine cell MC - 2 Machines: 2, 3, 5, 6 Parts: 5, 8, 10, 11

Functional machining facility Parts: 1, 4, 9

The University of Iowa

Intelligent Systems Laboratory

The University of Iowa

Intelligent Systems Laboratory

Three other alternative approaches

1. Process plans for parts 1, 4, and 9 might be modified, so that each of these parts could be machined in one of the two existing machine cells MC-1 or MC-2. 2. Designs of parts 1, 4, and 9 might be modified, so that the resulting process plans could fit the existing machine cell MC-1 or MC-2. 3. Parts 1, 4, and 9 might be manufactured in the two existing machine cells MC-1 and MC-2 without any changes of their process plans. This approach is applicable as long as the flow of parts among different machine cells (in the preceding case, MC-1 and MC-2) is relatively low.

The University of Iowa

Intelligent Systems Laboratory

Page 9
9