Module 8 – Line Balancing, Location and layout

1. Define the line balancing problem?

Given a set of tasks that are required to assemble a product along with precedence
relationships and task times, the problem is to allocate the tasks to minimum
number of workstations, satisfying the precedence constraints such that the time
taken in a workstation is within a given cycle time.

2. How do you solve the line balancing problem?

The line balancing problem can be formulated and solved using Integer
Programming. However since Integer programming problems are hard, we solve the
line balancing problem using heuristics. These heuristic algorithms do not guarantee
optimum solutions but give good solutions.

3. Name two measures for the goodness of a line balancing solution?

Two measures for the goodness of line balancing solutions are line efficiency and
smoothness index.

4. Define Line efficiency?

∑𝑡
Line efficiency 𝜂 = 𝑛×𝐶𝑇. Here Σt is the sum of task times, n is the number of
workstations and CT is the cycle time.

5. Define smoothness index?

Smoothness index SI = Σ(Tk – CT)2 where Tk is the total time in workstation k

6. Explain one heuristic algorithm specific to line balancing problem?

We explain the Helgeson-Birnie heuristic. Construct the network disgram and
compute the positional weight for each activity. The positional weight is the longest
path in the network starting from the activity. Rank the activities in the decreasing
order of positional weight. Create the first workstation and identify the activities that
can be allocated based on satisfying the precedence constraints. If more than one
activity can be allocated, choose the one with highest positional weight. If the cycle
time is exceeded, create a new workstation. Assign all the tasks and compute the
number of workstations and the cycle time in each.

7. Explain the p median problem for location?

The p-median problem is to group n points (things) into p groups where p is given. A
distance matrix is given among the n points. Out of these n points, p points are
chosen as medians and each the remaining point is grouped with the closest median.
This can be formulated as a binary integer programming problem and solved.

8. What is the difference between location and layout?

 Location problems address the place to locate a factory or facility while layout
problems address the relative location of facilities within a factory

9. Explain the quadratic assignment problem and its relationship to layout problem?
The quadratic assignment is a special type of an assignment problem where the
objective function is a quadratic function (product form) of the decision variables.
The constraints are the same as the assignment problem. The decision variables are
binary. It is related to the layout problem. If Xik = 1 when facility i is assigned to site k,
the objective function would be ΣΣΣΣ wijdklXikXjl. Where wij is the material movement
between facilities i and j and dkl is the distance between sites k and l. The constraints

10. Explain one quantitative method for layout?

The CRAFT algorithm is a quantitative model for layout. CRAFT stands for
Computerised Relative allocation of facilities technique. Here an initial layout is given
and the load distances are calculated. Through a pairwise exchange of facilities, the
load distances are minimized and a good layout is provided.

11. Explain one qualitative model for layout?

ALDEP is an example of a qualitative model for layout. The input is a preference
matrix made out of A, E, I, O, U. These represent the order of decreasing importance
of facilities located close to each other. We choose a facility randomly and place it.
Facilities with high importance are placed nearby. The layout is completed and
evaluated. Several layouts are created and the best can be found.

Problems

1. You are given a line balancing problem with the following data (A, ‐, 5), (B, A, 6), (C,
A, 4),(D, B and C, 6), (E, D, 8), (F, E, 7), Find a solution with minimum number of work
stationsfor T ≤ 12. Find the line efficiency and smoothness index?
One solution would be to use 4 workstations – WS1 has A and C with time = 9, WS2
has B and D with time = 12, WS3 has E with time = 8 and WS4 has F with time = 7.
Line efficiency = 36/(4x12) = 75%. SI = 9+0+16+25 = 50

2. You are given a line balancing problem with the following data (A, ‐, 6), (B, A, 3), (C,
A, 4),(D, B and C, 8), (E, D, 10), (F, D, 4), (G, E and F, 8), Find a solution with minimum
numberof work stations for T ≤ 12. Find the line efficiency and smoothness index?
One solution would be to use 4 workstations – WS1 has A and B with time = 9, WS2
has C and D with time = 12, WS3 has E with time = 10 and WS4 has F and G with time
= 12. Line efficiency = 43/(4x12) = 89.58%. SI = 9+0+4+0 = 13
3. Write the integer programming formulation of the line balancing problem? Why is
itadvantageous to start with a heuristic solution to the problem?
Let Si = 1 if station i is chosen and 0 otherwise
Let Xij = 1 if task j is assigned to station i and 0 otherwise

The objective function is to minimize the number of work stations. This is to

n

∑S i
Minimize i =1
n

∑X ij =1
Each task is assigned to only one work station. This is given by i =1

Each work station should have tasks such that the station time does not exceed T.
n

∑t j X ij ≤ TSi
This is given by j =1

∑𝑙𝑟=1 𝑋𝑗𝑟 ≥ 𝑋𝑘𝑙 ∀ 𝑙 = 1 𝑡𝑜 𝑛 ; X ij , S i = 0,1

It is advantageous to start with a heuristic solution to reduce the number of deciaion
variables and constraints.

4. Consider the five points whose distance matrix is given in Table 1

Table 1
12345
1 -- 20 28 24 16
2 20 -- 22 19 30
3 28 22 -- 32 20
4 24 19 32 -- 20
5 26 30 20 20 --
Assume that p = 2. Consider two out of the five points as medians and allocate the
rest of the pointsto the medians. Find the solution? Is it optimal? What can happen if
the number of points and number of groups increases?
Points 3 and 4 are medians. Point 1 is attached to 4, Point 2 is attached to 4, Point 5
is attached to 3. The solution is {1, 2, 4} and {3, 5}. Since we had p = 2 we assumed
that two points farthest from each other are the medians. This cannot be
generalized when p increases. The solution need not be optimum.

5. Consider five points (8,12), (6,11), (13, 14), (12, 14), (15, 17). Compute a distance
matrix and solve a p‐median problem for p = 2 and p = 3.
We compute rectangular distances. Points 2 and 5 are farthest. Point 1 is attached to
2, Point 3 to 5 and point 4 to 5. The groups are {1,2} and {3, 4, 5}. For p = 3, the seed
points are 2,5 and 3. The groups are {1, 2}, {3} and {4, 5}

6. Formulate a three stage supply chain where there are factories, warehouses,
distribution centres and customers. How many constraints and variables does the
formulation have?
 A supply chain with factories, warehouses and customers is given below:
Let Yi = 1 if plant i is opened
Let Wj = 1 if warehouse j is opened.
Let Xij be the quantity of the product transported from plant i to warehouse j.
Let Tjk be the quantity of the product transported from warehouse j to customer k.

The objective function is to minimize the total cost of location and allocation. This is to

m p m p p n
Minimize ∑
=i 1
fiYi + ∑ g jW j + ∑∑ Cij X ij + ∑∑ C jk T jk
=j 1 =i 1 =j 1 =j 1 =
k 1
Subject to
p

∑X
j =1
ij ≤ PY
i i ∀i

m n

∑ X ij ≥ ∑ T jk ∀j
=i 1 =k 1
p

∑T
j =1
jk ≥ d k ∀k

Yi , W j = 0,1
X ij , T jk ≥ 0
If we include distribution centres, we have location and allocation variables and
constraints. We also have material balance constraints.

7. If there is a fixed cost (fi) of making a point as a median, formulatethe problem as an

integer programming problem?
The p median formulation is Let Xjj = 1 if point j is a median and Xij = 1 if point i is attached
to median j. The objective is to
n n
Minimize ∑∑ d
=i 1 =j 1
ij X ij
n
Subject to ∑Xj =1
jj =p
n

∑X
i =1
ij =1

X ij ≤ X jj
Xij = 0,1
We can add a constraint ΣfiYi ≤ B if there is a budget restriction.

8. Consider the matrix of closeness of four facilities shown in Table 2

Table 2.
1234
1 -- A O I
2 -- E U
3 -- O
4 --
Provide a layout using the two qualitative algorithms explained in this chapter?
 One solution using ALDEP would start with 1, add 2 and then randomly choose 3 and
4. The arrangement could be 1-2-3 and 4. CORELAP would start with 2and then add 1
and 3 and 4. The arrangement could be 1-2-3-4.

9. Consider the work load matrix shown in Table 3 and the distance matrix shown in
Table 4
Table 3
1234
1 -- 4 5 3
2 3 -- 5 4
3 3 5 -- 4
4 5 3 5 --
Table 4
1234
1 -- 1 2 1
2 -- 1 2
3 -- 3
4 --
Find a layout that minimizes load distance?
Let aj be the site to which facility i is assigned. Consider [3 2 1 4] The load distance is
4 + 10 + 9 + 5 + 8 + 4 = 40. Consider [1 2 3 4]. The load distance is 4 + 10 + 3 + 5 + 8 +
12 = 42 and so on. The solution [3 2 1 4] is a good solution. Since the matrices are
symmetric, the actual load distance is 80.

10. Explain the principle behind the working of the CRAFT algorithm
The CRAFT algorithm is a quantitative model for layout. CRAFT stands for
Computerised Relative allocation of facilities technique. Here an initial layout is given
and the load distances are calculated. Through a pair wise exchange of facilities, the
load distances are minimized and a good layout is provided. The exchange is
considered only if the facilities share a common border so that the smaller area
would be put inside the larger area. The centroids are used to compute the
distances.