You are on page 1of 4

Excel Modeling for extended Johnson Rule of sequencing

Johnson Rule as of 1954 is used for optimal solution for sequencing n jobs on two machines.
Extended Johnson rule for more than two machines club the problem and make it two machine
problem to again carry out same algorithm.
Johnson rule algorithm has a specified rule of allocation of job on machine. Least processing
time job on first machine will go first in sequence, Least processing time job on second machine
will go last in sequence and hence on. It looks easy for small no of job and only two machines
but becomes time consuming and complex n jobs and n machine.
Here I have attempt to develop a Mix integer linear programming model for Johnson rule and
extended Johnson rule on more than two machines.
2 Machine n job sequencing:
Basic assumptions are:
1.) A job must follow the sequence over machine i.e. before moving to second machine it
has to go through 1
st
machine
2.) Job will have same sequence on 1
st
as well as on 2
nd
machine
Optimizing Equations:
In sequencing problem we can get the optimal solution by minimizing either of the below three
attribute. Deciding which one to minimize is solely depends on the management i.e. on what
they want to minimize.
1.) Average flow time: Average of the cumulative sum of the jobs flow on the two machines.

For machine 1 Cumulative flow time:

( ) ( )
(

+ =

=

T
f
P T
f
t i
n
i
it t 1
1
1
1
1
o


For machine 2 cumulative flow time:
( ) ( ) ( ) ( ) ( )
(

+ + =

=

T
f
T
f
T
f
P T
f
t t t i
n
i
it t 1
2 1
1
2
2
1
2
, 0 max
o

Where,

n is the number of job

o
it
= 1 if job selected for process in period t, 0 otherwise
Pi
1
= Process time of i
th
job on machine 1
Pi
2
= Process time of i
th
job on machine 2

( ) ( ) ( )
T
f
T
f
t t 1
2 1
, 0 max

= Waiting time of machine 2




Equation to optimize:


( ) ( ) | | | |
T
f
T
f
N N
avg
2 1
min +


2.) Make-span time: Time taken to finish all job on all machines. It is the cumulative sum of
all job flow time on machine 2.

Equation to optimize:
( ) ( ) ( ) ( )
(

+ +

=

T
f
T
f
T
f
P t t t i
n
i
it
Min
1
2 1
1
2
2
1
, 0 max
o


3.) Ideal Time: It is the time machine 2 and machine 1 sit ideal.

Equation to optimize:
( ) ( ) ( ) ( )|
.
|

\
|
+ +

=

=
T
f
T
f
T
f
T
f
P N N
N
t
t t i
n
i
i
Min
1 2
1
1
2 1
1
1
1
, 0 max
o

Here one thing is to be noted that scheduling sequence could be different for different optimizing
equation used. Sequencing could also be different from Johnson rule solution.
Constraints:
Constraints are pretty simple for this. They are:
1.)
o
it
should be binary
2.)
1
1
=

=
n
t
it
o ; Job n should be scheduled exactly 1 time
3.)
1
1
=

=
n
i
it
o ; No more than 1 job can be scheduled in one period

n Job n machine problem:
n job n machine problem are pretty similar to 2 machine n job problem. Constraints are
absolutely same. Only optimizing equation takes an extended for.
Optimizing Equations:
1.) Average flow time: Average of the cumulative sum of the jobs flow on the two machines.

For machine 1 Cumulative flow time:

( ) ( )
(

+ =

=

T
f
P T
f
t i
n
i
it t 1
1
1
1
1
o


For machine n cumulative flow time:
( ) ( ) ( ) ( ) ( )
(

+ + =

T
f
T
f
T
f
P T
f
t
N
t
N
t
N
N
i
n
i
it t
N
1
1
1
1
, 0 max
o

Where,

n is the number of job
N is the machine number
o
it
= 1 if job selected for process in period t, 0 otherwise
Pi
1
= Process time of i
th
job on machine 1
P
n
i
= Process time of i
th
job on machine n

( ) ( ) ( )
T
f
T
f
t
N
t
N
1
1
, 0 max


= Waiting time of machine N


Equation to optimize:


( ) ( ) | | | |
T
f
T
f
T T
avg
2 1
min +


4.) Make-span time: Time taken to finish all job on all machines. It is the cumulative sum of
all job flow time on machine n.

Equation to optimize:
( ) ( ) ( ) ( )
(

+ +

T
f
T
f
T
f
P t
N
t
N
t
N
N
i
n
i
it
Min
1
1
1
1
, 0 max
o


5.) Ideal Time: It is the time machine 2 and machine 1 sit ideal.

Equation to optimize:
( ) ( ) ( ) ( )|
.
|

\
|
+ +

=

=

T
f
T
f
T
f
T
f
P T
N
T
N
T
t
t
N
t
N
i
n
i
i
Min
1
1
1
1
1
1
1
, 0 max
o


Here one thing is to be noted that scheduling sequence could be different for different optimizing
equation used. Sequencing could also be different from Johnson rule solution.
An excel example is attached with this article. I have taken a hypothetical example to illustrate
how to use MILP for sequencing n job. Sheet 1 have an example of n Job on 2 machine and
sheet 2 have an example of n job n machine sequencing.
Hope you have enjoyed the article and gained out of it. Please feel free for any comment and
clarification.

You might also like