Facilities Planning and Warehouse Operations

INDU 6251 – Fall 2023

Flow, Space and Activity Relationships

Makbule Kandakoglu, PhD

Some of the slides are adapted from the course textbook and the lecture notes provided by Prof. Andrea Schiffauerova and Prof. Ali Akgunduz
Outline
• Departmental planning
• Manufacturing cells
• Clustering algorithms for cell formation

2
Departmental Planning
• Production planning departments are collections of workstations to be grouped
together during the facilities layout process

• Combining workstations that perform similar functions:

• Similar products or components
• Similar processes

• Classification of layouts based on product volume-variety:

• Product layout (flow shop)
• Fixed product layout
• Group layout
• Process layout (job shop)

3
Alternative Types of Layouts

Product layout (flow shop) Group layout

Fixed product layout

Process layout (Job shop)

4
Alternative Types of Layouts

5
Alternative Types of Layouts

6
Alternative Types of Layouts

7
Alternative Types of Layouts
Production volume and product variety determines type of layout

8
Group Technology – Cellular Manufacturing
• Group technology (product family) departments aggregate medium volume-variety parts into
families based on similar manufacturing operations and design attributes.
• Production workstations and equipment are arranged in a sequence that supports a smooth
flow of materials and components through the production process with minimal transport or
delay.

Similar design attributes, Different design attributes,

different manufacturing requirements similar manufacturing requirements

9
Group Technology – Cellular Manufacturing
Goal
• Create manufacturing cells
• Perform the machining task for a given product in a single cell
• Minimize the transportation between cells

10
Group Technology – Cellular Manufacturing
• Manufacturing cells group machines, employees, materials, tooling and
material handling and storage equipment to produce families of parts.

• Benefits of cell manufacturing:

• Reduction: inventories, space, paperwork, equipment, transportation, etc.
• Simplification: communication, handling, scheduling, etc.
• Improvement: productivity, flexibility, quality, customer satisfaction, etc.

11
Group Technology – Cellular Manufacturing
2
Before GT 1 3 P1 P1: 6 → 5 → 4 → 3

P2: 6 → 5 → 2 → 3
P2
P3: 6 → 5 → 1 → 3

P3
4 5 6

5 1 P3
6 3
After GT 5 2 P2

5 4 P1

13
Group Technology – Cellular Manufacturing

Before GT

Cell 1

After GT

Cell 2

14
Group Technology – Cellular Manufacturing

Inventories, space, simplification,

communication, scheduling
15
Cellular Manufacturing
• Successful implementation requires addressing selection, cell design, cell operation
and cell control issues

• Manufacturing cell forming:

• Classification
• Production flow analysis
• Clustering methodologies
• Heuristic procedures
• Mathematical models

• Cell forming is not often the responsibility of a facility planner

16
Clustering Methodologies
• Group parts together so that they can be
processed as a family

• Links parts and machines in machine-part

matrix

Algorithms:
• Direct clustering algorithm (DCA)
Machine-part matrix
• Binary ordering algorithm (BOA)
• Cluster identification algorithm (CIA)

17
Direct Clustering Algorithm (DCA)
1. Form machine-part matrix
2. Sum the 1s in each column & row
3. Order the rows in descending order
4. Order the columns in ascending order
5. Sort the columns (1 in the first row moves left, then in the second row, etc.)
6. Sort the rows (1in the first column moves upward, then in the second column,
etc.)
7. Form sells

18
Direct Clustering Algorithm (DCA) – An Example

1. Form machine-part 2. Sum the 1s in each 3. Order the rows in

matrix column & row descending order
4. Order the columns in
ascending order

19
Direct Clustering Algorithm (DCA) – An Example

5. Sort the columns

(1 in the first row moves the column to the left, then 1 in the second row, etc.)

20
Direct Clustering Algorithm (DCA) – An Example

6. Sort the rows

(1 in the first column moves the row upward, then 1 in the second column, etc.)

21
Direct Clustering Algorithm (DCA) – An Example

7. Form cells

Cell #1:
Machines 2, 4 and 5

Cell #2:
Machines 1 and 3

22
Direct Clustering Algorithm (DCA) – Bottleneck Machines

DCA

Machine 2 which is needed for parts #3 and #5 creates conflicts!

23
Direct Clustering Algorithm (DCA) – Bottleneck Machines

What can we do?

24
Direct Clustering Algorithm (DCA) – Bottleneck Machines
Possible solutions:
• Locate the bottleneck machines close to each other
• in different cells
• at the boundary between cells
• Redesign the parts
• Outsource the parts
• Duplicate machines

25
Direct Clustering Algorithm (DCA) – Bottleneck Machines
Duplicate machines

Duplicated machine 2 Duplicated machine 3

26
Direct Clustering Algorithm (DCA) – Bottleneck Machines

DCA

27
Direct Clustering Algorithm (DCA) – Bottleneck Machines
• Create only 2 cells

• Place the bottleneck

machines at the
boundary of the cells

• Duplicate the
bottleneck machines

Can the problem be solved by redesigning or outsourcing?

28
Binary Ordering Algorithm (BOA)
Binary ordering algorithm (Rank ordering algorithm) considers the rows and
columns as binary strings

Procedure:
1. Compute the decimal equivalent of the binary strings for rows
2. Reorder the rows in decreasing order of their binary value
3. Compute the decimal equivalent of the binary strings for columns
4. Reorder the columns in decreasing order of their binary value
5. If the machine-part matrix is unchanged, then stop, else repeat

29
Binary Ordering Algorithm – An Example
• A manufacturing firm produces 10 different products using 5 different machines.
Following is the manufacturing sequence of each product.

• Create manufacturing cells where most of Parts Machining Requirements

the machining of the selected part are 1 M1 M3 M5

completed in the cell 2 M1 M2 M4 M5
3 M1 M2 M4 M5
• You may use more than one of each kind
4 M1 M2 M4 M5
of machines
5 M1 M3 M5
6 M3 M5
7 M1 M3 M5
8 M1 M4 M5
9 M1 M2 M4
10 M1 M2 M4

30
Binary Ordering Algorithm – An Example
Machine-part matrix

31
Binary Ordering Algorithm – An Example
1. Assign binary weights from right to left for components and calculate the decimal
equivalent for each row (machine)

29 + 28 + 27 + 26 + 25 + 23 + 22 + 21 + 20 = 1007

32
Binary Ordering Algorithm – An Example
2. Rank the machines according to their decimal equivalent computed in the previous
step
3. Assign binary weights from bottom up for machines and calculate the decimal
equivalent for each column (part)

24 + 23 + 22 = 28 33
Binary Ordering Algorithm – An Example
4. Reorder the columns in decreasing order of their binary value
5. If the machine-part matrix is unchanged, then stop, else repeat

For those components where the required machines could not be grouped together:
• Either multiple machines will be used or
• Components will be allowed to travel between cells

34
Cluster Identification Algorithm (CIA)
1. Select any row and cross it
2. For each crossed 1 make a vertical line
3. For each crossed 1 make a horizontal line
4. Repeat until all the 1s are crossed by a vertical line or by a horizontal line
5. Form a cell from all the machines and components which were crossed
6. Remove all the crossed elements (machines and components) and start
again

35
Cluster Identification Algorithm – An Example
1. Select any row and cross it

36
Cluster Identification Algorithm – An Example
2. For each crossed 1 make a vertical line

37
Cluster Identification Algorithm – An Example
3. For each crossed 1 make a horizontal line

38
Cluster Identification Algorithm – An Example
4. Repeat until all the 1s are crossed by a vertical line or by a horizontal line

39
Cluster Identification Algorithm – An Example
5. Form a cell from all the machines and components which were crossed

• First cell is identified!

• Cell #1 will produce parts P2, P3, P5 and P8 with Machines M1, M5 and M7

40
Cluster Identification Algorithm – An Example
6. Remove all the crossed elements (machines and components) and start again

• Cell #2 will produce parts P1 and P6 with Machines M2 and M4

41
Cluster Identification Algorithm – An Example
6. Remove all the crossed elements (machines and components) and start again

• Cell #3 will produce parts P4 and P7 with Machines M3 and M6

42
Cluster Identification Algorithm – An Example
Three resulting cells

43
Cellular Manufacturing
• In the real-world cases the solution will rarely be straightforward
• Besides the machine-part matrix, some other factors that may affect the cell
formation:
• Machine Cost (Pm)
• Manufacturing Times (Tij)
• Capacity of Machines (Ri)
• Inter-cell material handling cost (Hnm)
• Demand (Di)
• Number of Machine Assignments in Cells (Min)
• Assignment of jobs for Cells (Jicm)
• Travelling for machining (Vim)
• …

44
Next Lecture
• Planning for personnel
• Material handling

45