MDP408a

Production & Operations

Management (Facilities Planning)
Lecture #13
Facility Layout Planning (cont.)
 Today’s lecture
• Space requirements
• Space relationship diagram
• Systematic Layout Procedure (SLP)
• Layout optimization
• Space requirements

Space Requirements
• Perhaps the most difficult aspect of facility plan
• Large amounts of uncertainty
– Technology changes
– Demand forecasts
– Changing product mix
– Parkinson’s Law – things will fill available space
– New designs (pull systems, decentralized storage, etc)
• Generally use “bottom-up” approach
– Define workstation spaces
– Define departmental spaces (with aisle allowances)
• Space requirements

Workstation Space
Sample Lathe Workstation
• Space requirements

Departmental Space
• Space requirements
Departmental Service Requirements
Sheet
• Space requirements

Aisle Planning
• Objective: Promote Flow
• Two Types:
– Departmental Aisles (previous slides)
– Main Aisles
• Tradeoff
– Congestion versus Wasted Space
• Principles:
– Avoid curves and non-right-angle intersections
– Aisles should be straight and lead to doors
– Avoid aisles along outer wall of bldg. Unless for entry/exit
– Don’t forget to consider column spacing!
• Space requirements

Some Recommended Aisle Widths

• Space requirements

Personnel Space Requirements

Greatly a function of owner philosophy!
1. Employee/Facility Interface
2. Restrooms
3. Food Services
4. Health Services
5. Barrier Free Compliance
6. Office Planning
Office Space Planning Example
 Space Relationship Diagram
• Combine space requirements with relationship
information to develop space relationship diagram
• Expand sizes of relationship chart nodes to reflect
approximate shape and size of areas.
• Generate multiple alternatives
• Same rules apply as in the development of the
original relationship diagram.
• Measuring flow

Activity Relationship Chart Example

The Relationship Diagram
Relationship Diagram and Space
Requirements
Space Relationship Diagram
Muther’s systematic layout planning (SLP)
procedure

• Developed by Muther in 1973 to construct

new layouts.
Muther’s Systematic Layout Planning (SLP)
Muther’s Systematic Layout Planning (SLP)
– Contd.
 Algorithmic Approaches (i)
• SLP is “informal”
• Does not provide a formal procedure or
algorithm for critical steps
• Algorithm: a precise rule (or set of rules)
specifying how to solve some problem
– Has an objective function
– Amenable to computer implementation
 Algorithmic Approaches (ii)
• Types of input
– Qualitative “flow” data (i.e. relationship chart)
– Quantitative data (i.e. from-to chart, flow-between)
– Both
• Classification of algorithms
– Based on objective function
• Distance based
• Adjacency based
– Format of layout representation
• Discrete representation
• Continuous representation
– Primary function
• Improvement
• Construction
 Layout evaluation functions
• Layout development criteria
– When constructing a layout, an algorithm
implemented on a computer needs specific criteria
to compare alternative layouts
– These criteria may differ depending on the form of
input data characterizing flow/relationships
– The criteria must be computable (quantitative) and
is referred to as the objective function of the layout
problem
 Layout evaluation functions
(Distance-based objective)
• From-to chart as input data
– The objective will be to construct a layout to
minimize required movement costs over a specific
time period
Notation:
m = The number of departments in the layout.
fij = The department i to department j flow (from-to data).
cij = The cost of moving a load one distance unit from department
i to department j.
dij = The distance from department i to department j.
z = Total movement cost per time period (the objective function value).
 Layout evaluation functions
(Distance-based objective)
• The objective function can be expressed as:

m m
z =  f ij cij d ij
i =1 j =1

• This is called a distance-based objective function

• The fij and cij are input data
• If the cij= 1 for all i and j, z= total travel distance per time
period
 Distance Measure
• The dij’s are normally rectilinear centroid to centroid
distance
• The dij’s change as a function of department location
and shape
• Rectilinear distance’s

If the centroid of department1 is located at ( x1 , y1 )

and the centroid of department 2 is located at ( x2 , y2 )

then d12 =| x1 − x2 | + | y1 − y2 |
 Centroids
• Centroid
– Center of gravity or mass
• Find the centroid
y
40
35
30
25
20
15
10
5
0
0 10 20 30 40 50 x
 Layout evaluation functions
(adjacency-based objective)
• Activity relationship chart as input
– The activity relationship between departments i and
j must first be quantified to get fij, which represents
the relationship strength on some numerical scale
• One possible objective function:

m m
z =  f ij xij
i =1 j =1

1 if depts. i and j are adjacent

where xij = 
0 otherwise
 Layout evaluation functions
(adjacency-based objective)
• This is called an adjacency based objective
• Since the numerical scale is arbitrary, another
objective function is often used
– Called an efficiency rating
– It is a relative measure of a layout with respect to
the best (sometimes not possible) objective
function value
 Layout evaluation functions
(adjacency-based objective)
m m

 f x
ij ij
• No negative fij values z=
i =1 j =1
m m

 f
i =1 j =1
ij

 fij xij −  fij (1 − xij )

z = i , jF i , jF


i , jF
fij − 
i , jF
fij

• Negative fij values

F = The set of dept. pairs with fij  0
F = The set of dept. pairs with fij  0
 Classification of Algorithms - Primary
functions
▪ Construction-based
– Develop ‘from scratch' and progressively build layout
– Example
• ALDEP (Automated Layout Design Program)
• CORELAP (Computerized Relationship Layout Planning)
• Improvement-based
– Start with an initial layout and try to improve it through
incremental changes
– Example
• Pair-wise exchange (TWOWAYX)
• CRAFT (Computerized Relative Allocation of Facilities Technique)
 Classification of Algorithms
- Layout Representation Format
• Layout representation format
– Discrete
• The area of each department is rounded off to the nearest integer number of grids.
• A smaller grid size yields a finer resolution and gives more flexibility in department
shapes, but
• Results in a larger number of grids which complicates computations. Why?
– Continuous
• Does not use a grid
• More flexible but more difficult to use
• Usually limited to rectangular building and departments
Examples of Split and Unsplit departments
 “Optimizing” a Layout
• Finding the optimal layout that minimizes either of the objective
functions described (distance-based or adjacency-based) is a very
difficult problem (computationally)
• General methods for exploring the space of layouts have been
applied
– Graph-based method
• A construction algorithm used to build a new layout
– Pairwise exchange method
• A general approach used to improve an existing layout
– Other methods
• Metaheuristics: Genetic Algorithm, Tabu Search, Simulated
Annealing…etc.
 Graph-based Method
• Often applied when using an adjacency-based objective
• A method that uses a graph (nodes connected by arcs) to
represent departments and their adjacencies to other
departments
– Departments = nodes
– If an arc connects two departments, they are adjacent
• Arcs have values that represent adjacency relationships
– Finding the best layout amounts to finding the best graph
 Graph-based Method
• Procedure for generating an adjacency graph
1. Find the highest valued adjacency relationship as the
starting nodes connected by an arc (break ties randomly)
2. The third department selected will have the largest total
relationship value with the two departments selected in
step 1
3. The next departments to enter are selected one at a time
based on the highest total relationship values added when
the department is placed on the “face” of the existing
graph
▪ A face is a region bounded by arcs of the graph
Example

Step 1

3 4
20
Graph-based Method
 Graph-based Method (2)

• A department block layout still needs to be

constructed from the graph
Graph-based Method