Applications of the PMP. Cell Formation in Group Technology

44
Applications of the PMP Cell Formation in Group Technology

description

AACIMP 2010 Summer School lecture by Dmitry Krushinsky. "Applied Mathematics" stream. "The p-Median Problem and Its Applications" course. Part 5. More info at http://summerschool.ssa.org.ua

Transcript of Applications of the PMP. Cell Formation in Group Technology

Page 1: Applications of the PMP. Cell Formation in Group Technology

Applications of the PMP

Cell Formation in Group

Technology

Page 2: Applications of the PMP. Cell Formation in Group Technology

Outline

• Introduction

• Brief overview of models

• PMP approach to CF

• Examples

Page 3: Applications of the PMP. Cell Formation in Group Technology

Group technology

• A paradigm in industrial engineering

suggesting structural decomposition of a

manufacturing system into smaller

subsystems.

Page 4: Applications of the PMP. Cell Formation in Group Technology

Group technology: advantages

• Smaller systems are easier to manage

(e.g. scheduling)

• Better plant layout:

– shorter in travelling distances (up to 95%)

– less intersecting routes

Page 5: Applications of the PMP. Cell Formation in Group Technology

Cell formation (CF)

• Grouping machines into manufacturing

cells …

• … and parts into product families …

• … such that each family is produced

(mainly) within one cell

Page 6: Applications of the PMP. Cell Formation in Group Technology

Cell formation

• Cell Formation becomes possible by

exploiting similarities in the manufacturing

processes for different parts and increases

the throughput of the manufacturing

system without sacrificing the products

quality.

Page 7: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 8: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 9: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 10: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 11: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 12: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 13: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Page 14: Applications of the PMP. Cell Formation in Group Technology

ExampleDrilling and cutting

Thermal processing

Cell 1

Cell 2

Cell 1

Cell 2

Page 15: Applications of the PMP. Cell Formation in Group Technology

Machine-part incidence matrix

machin

es

parts

.

1

.

.

1

.

1

.

.

1

1

.

.

1

.

.

.

1

1

.

1

.

1

.

.

.

.

1

1

1

Page 16: Applications of the PMP. Cell Formation in Group Technology

Machine-part incidence matrix

.1.1..

111...

1.11..

....11

1...11

.1..1.

...1.1

111...

1.1.1.

1..1.1

5

4

3

2

1

5

3

2

4

1

654321 654231

Page 17: Applications of the PMP. Cell Formation in Group Technology

Performance measures

1111...1....

1.11........

....11111...

....1.111...

.1..111.1...

.........111

..1......111

.........111 - exceptions

- voids

ne – number

of exceptions

nv – number

of voids

no – total

number of 1s

r

m

machin

es

parts

Page 18: Applications of the PMP. Cell Formation in Group Technology

Performance measures

number of cellsm1

exceptions, ne

voids, nv

Page 19: Applications of the PMP. Cell Formation in Group Technology

Performance measures

• ne

• ne + nv

• many others

nenvnorm

nvnorm

nvneno

neno)1(

factor g weightin ]1;0[

Page 20: Applications of the PMP. Cell Formation in Group Technology

Problem size reduction

machine-part

incidence matrix

mxr

machine-machine

similarity matrix

mxm

machine-machine

similarity measure

Similarity measures:

.1..1.

...1.1

111...

1.1.1.

1..1.1

M5

M4

M3

M2

M1Each machine is characterized by a

vector in r-dimensional space

similarity any computable metrics

r parts

Dissimilarity: constjisjid ),(),(

),( jis

Page 21: Applications of the PMP. Cell Formation in Group Technology

Problem size reduction

machine-part

incidence matrix

mxr

machine-machine

incidence matrix

mxm

machine-machine

similarity measure

Wei and Kern similarity measure:

r

k

jkik aajis1

),(),(

jkik

jkik

jkik

jkik

aa

aa

aar

aa

if,0

0 if,1

1 if,1

),(

r – number of parts,

m – number of machines

Page 22: Applications of the PMP. Cell Formation in Group Technology

CF: existing approaches

• Clustering based on energy functions: BEA, ROC, MODROC, DCA, …

• Similarity based hierarchical clustering: SLC, CLC, ALC, LCC, …

• Fuzzy logic methods

• Genetic algorithms and simulated annealing

• Neural networks: backpropagation learning, competitive learning, adaptive resonance theory (ART1), self-organizing maps, …

• Graph partitioning

• Integer Linear Programming

Page 23: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: BEA

.1..1.

...1.1

111...

1.1.1.

1..1.1

BEA = bond energy analysis

Goal: minimize the length of the border

11....

1.11..

.111..

....11

...111

• equivalent to the Quadratic Cost Assignment Problem

• only partial solution

Page 24: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: hierarchical

clustering

SLC/CLC/ALC = single/complete/average linkage clustering

Algorithm:

• start with each cluster containing one machine

• at each step connect two most similar clusters

Page 25: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: hierarchical

clustering

SLC/CLC/ALC = single/complete/average linkage clustering

Algorithm:

• start with each cluster containing one machine

• at each step connect two most similar clusters

Page 26: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: hierarchical

clustering

SLC/CLC/ALC = single/complete/average linkage clustering

Algorithm:

• start with each cluster containing one machine

• at each step connect two most similar clusters

Page 27: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: hierarchical

clustering

SLC/CLC/ALC = single/complete/average linkage clustering

Algorithm:

• start with each cluster containing one machine

• at each step connect two most similar clusters

Page 28: Applications of the PMP. Cell Formation in Group Technology

Existing approaches: hierarchical

clustering

SLC/CLC/ALC = single/complete/average linkage clustering

Algorithm:

• start with each cluster containing one machine

• at each step connect two most similar clusters

Equivalent to

the minimum

spanning tree

problem

(MST)

Page 29: Applications of the PMP. Cell Formation in Group Technology

P-Median approach

Goal:

• select p “central” machines – representatives of p

cells

• assign all other machines to cells...

• ... such that the sum of dissimilarities is minimum

Page 30: Applications of the PMP. Cell Formation in Group Technology

P-Median approach

Goal:

• select p “central” machines – representatives of p

cells

• assign all other machines to cells...

• ... such that the sum of dissimilarities is minimum

p = 2

Page 31: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

parts

ma

ch

ine

s

1 2 3 4 5 6

5

4

3

2

1

Machine-part

incidence matrix

Goal: group machines into clusters

(manufacturing cells) such as to minimize

intercell communication.

r

k

jkik aarrjid

1

),()1(),(

Wei and Kern’s “commonality score”

jkik

jkik

jkik

jkik

aa

aa

aar

aa

if,0

0 if,1

1 if,1

),(

r – number of parts, m – number of machines

m = 4, r = 5

.1..1.

...1.1

111...

1.1.1.

1..1.1

Page 32: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

Cost matrix for the PMP

is a machine-machine

dissimilarity matrix:

),(: jidcij

.1..1.

...1.1

111...

1.1.1.

1..1.1

1628232329

2816292917

2329121824

2329181224

2917242412

r

k

jkik aarrjid

1

),()1(),(

Page 33: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

1628232329

2816292917

2329121824

2329181224

2917242412

C

5432532525

5421541414

5321532323

5321532322

4321421411

150716

0111116

515612

515612

507512)(

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyyBC y

5325413241543212, 7110187166568)( yyyyyyyyyyyyyyyB pC y

Page 34: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

1628232329

2816292917

2329121824

2329181224

2917242412

C

5432532525

5421541414

5321532323

5321532322

4321421411

150716

0111116

515612

515612

507512)(

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyyBC y

5325413241543212, 7110187166568)( yyyyyyyyyyyyyyyB pC y

New variables:

5329

5418

327

416

yyyz

yyyz

yyz

yyz

Additional constraints:

1or 2

1or 2

1

1

5795329

5685418

327

416

yzzyyyz

yzzyyyz

yyz

yyz

Page 35: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

min7110187166568 987654321 zzzzyyyyy

s.t.

5...1 },1,0{

9...6 ,0

1

1

1

1

25

579

568

327

416

54321

iy

iz

yzz

yzz

yyz

yyz

yyyyy

i

i1

1

1

0

0

*y

MBpBM formulation

Page 36: Applications of the PMP. Cell Formation in Group Technology

Example 1: functional grouping

1

1

1

0

0

*y

.1.1..

111...

1.11..

....11

1...11

654231

5

3

2

4

1

Page 37: Applications of the PMP. Cell Formation in Group Technology

Example 2: workforce expenses

Machine-worker

incidence matrix

11001000

00101100

10010110

00100011

01010001

workers

ma

ch

ine

s

1 2 3 4 5 6 7 8

5

4

3

2

1

Goal: group machines into clusters

(manufacturing cells) such that:

1) every worker is able to operate every

machine in his cell and cost of additional cross-

training is minimized;

2) if a worker can operate a machine that is not

in his cell then he can ask for additional

payment for his skills; we would like to minimize

such overpayment.

Dissimilarity measure for machines

machines theofeither operatecan that workersofnumber

and machinesboth operatecan that workersofnumber ),(

jijid

Page 38: Applications of the PMP. Cell Formation in Group Technology

Example 2: workforce expensesCost matrix for the PMP

is a machine-machine

dissimilarity matrix:

080.083.000.180.0

80.0083.080.000.1

83.083.0083.083.0

00.180.083.0080.0

80.000.183.080.00

machines

ma

ch

ine

s

),(: jidcij

11001000

00101100

10010110

00100011

01010001

workers

ma

ch

ine

s

5

4

3

2

1

87654321

Page 39: Applications of the PMP. Cell Formation in Group Technology

Example 2: workforce expenses

080.083.000.180.0

80.0083.080.000.1

83.083.0083.083.0

00.180.083.0080.0

80.000.183.080.00

C

5431541515

5432542424

4321321313

4321421212

5321521211

17.003.0080.0

17.003.0080.0

00083.0

17.003.0080.0

17.003.0080.0)(

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyy

yyyyyyyyyyBC y

543213, 8.08.083.08.08.0)( yyyyyB pC y

The objective is already a linear function !

Page 40: Applications of the PMP. Cell Formation in Group Technology

Example 2: workforce expensesMBpBM formulation

0

0

0

1

1

*y

5,...,1 },1,0{

2

s.t.

min8.08.083.08.08.0

54321

54321

iy

yyyyy

yyyyy

i

10010100

10101000

01010001

01100010

00011111

workers

ma

ch

ine

s

1

5

2

4

3

76418532 1 addtional training

7 redundant skills

Page 41: Applications of the PMP. Cell Formation in Group Technology

Example 3: from [BhaSad] (2010)*

* R. Bhatnagar, V. Saddikuti, Models for cellular manufacturing systems

design: matching processing requirements and operator capabilities,

Journal of the Operational Research Society, 61, 2010, pp. 827-839.

105 parts

46 m

achin

es

(uncapacitated)

functional grouping

105 parts

46 m

achin

es

grouping efficiency:

[BhaSad] 90.98%

our result 95.20%

(solved within 1 sec.)

Page 42: Applications of the PMP. Cell Formation in Group Technology

Future research directions

• Additional real-life constraints

– capacities

– workload

• Additional real-life factors

– operational sequences

– processing and setup times

Page 43: Applications of the PMP. Cell Formation in Group Technology

Conclusions

• An efficient model for CF:

– low computing times

– high quality solutions

Page 44: Applications of the PMP. Cell Formation in Group Technology

Conclusions

• An efficient model for CF:

– low computing times

– high quality solutions

• BUT: all models in literature (including our)

are heuristics from the CF perspective

• exact model – MINpCUT