Sunteți pe pagina 1din 37

Production Scheduling

• ‘n’ jobs to be processed – each has setup time, processing time,


and a due date.
• To be completed – each job is to be processed at several machines.
• Sequencing the jobs on the machines to optimize a certain
performance criterion.
List of performance criteria – to be optimized
• Mean flow time – It represents the average time each job spends in the shop

• Idle times of machines –

The period during which a machine is not being used. This is similar in
effect to machine-down time though it may be caused by lack of work rather than
by a fault in the machine

• Mean earliness of jobs – Job is completed before its due date

• Mean tardiness of jobs – Job is completed after its due date

• Mean queue time – The time before work that adds value to a product

• Mean number of jobs in the system - Measures amount of work-in-progress;


measures responsiveness and work-in-process inventory

• Percentage of jobs late


Classification of Scheduling
Single machine scheduling
• n jobs with same single operation on each of the jobs.

Flow shop scheduling


• n jobs with m operations on each of the jobs.
• Same process sequences.

Job shop scheduling


• n jobs with m operations on each of the jobs.
• Process sequences will be different from each other.
Basic data to describe jobs
Processing time (tj)
• Time required to process job ‘j’.
• It include both actual processing time and setup time.
Ready time (rj)
• Job ‘j’ is available for processing.
• Difference between arrival time of job and time at job taken for
processing.
Due date (dj)
• Time at which job ‘j’ is to be completed.
Completion time (Cj)
• Time at which job ‘j’ is completed in sequence.
• Performance measures for evaluating schedules are usually function
of job completion time.
Flow time(Fj)
• Amount of time job ‘j’ spends in the system.
• Indicates the waiting time of jobs in the system.
Fj = cj – rj
Lateness(Lj)
Lj = cj – dj
• Gives an idea about conformity of the jobs in a schedule to a given
set of due dates of the jobs.
• Positive lateness (completed after due date), it is measure of poor
service.
• Negative lateness (completed before due date), it is measure of
better service.
Tardiness(Tj)
• Fails to meet its due date, or zero, otherwise
Tj = max {0, Lj}
Makespan

• Total length of the schedule (that is, all the jobs have finished
processing).

• Amount of time, from start to finish for completing a set of


jobs, i.e. the maximum completion time of all jobs.

• The time difference between start and end time of a schedule.

• Completion time of the last job to leave the system.


Measures of Performance
• Mean flow time
𝟏 𝒏
Gives the average time each job spends in the shop, F̅= 𝒋=𝟏 Fj
𝒏

𝟏 𝒏
• Mean Tardiness 𝑻=𝒏 𝒋=𝟏 Tj

• Maximum flow time


Longest of the flow times of the jobs, Fmax =𝒎𝒂𝒙 {Fj }
𝟏≤𝒋≤𝒏

• Maximum tardiness
Longest of the tardiness of the jobs, Tmax =𝒎𝒂𝒙 {Tj }
𝟏≤𝒋≤𝒏

• Number of Tardy jobs


Number of jobs that are completed after their due dates,
NT = 𝒏𝒋=𝟏 f(Tj)
where, f(Tj) = 1, if Tj > 0, and
f(Tj) = 0, otherwise
Note: Tj = max {0, (cj – dj)}
Problem
Two jobs A and B, are required to be scheduled on 2 m/c, M1
and M2. Each job is processed first on M1, then on M2. The
processing times of these jobs are as follows. Find the
sequence of jobs that minimizes makespan.

Job MC 1 MC 2

A 8 5

B 7 10
Solution

Job A is scheduled first, Makespan – 25 Gantt chart

MC 1 Job A Job B

MC 2 Idle Job A Idle Job B


0 8 Time 13 15 25

Job B is scheduled first, Makespan - 22

MC 1 Job B Job A

MC 2 Idle Job B Job A


0 7 Time 15 17 22
n Jobs, one Machine (Single Machine Scheduling Problem)
• n jobs to be processed by one machine, all jobs go through this machine.
• Many practical applications – machine, a robot, a manufacturing cell etc.
• Time required to process a job varies according to job requirement.
• Sequence these jobs – mean flow time (MFT) of jobs is minimized.
1 𝑛
MFT = 𝑖=1 Fi
𝑛

1 𝑛 1 𝑛
MFT = 𝑖=1 (Ci - ri ) MFT = 𝑖=1 Ci
𝑛 𝑛

Ci – Completion time of job i,

ri – Ready time, (Single operation jobs is available for processing at time zero), ri = 0
n = No. of jobs to be processed.
Problem
Consider the following single machine scheduling problem.
Find the optimal sequence which will minimize the mean flow
time and also obtain the minimum mean flow time.
Job Processing Time
(hrs)
J1 7

J2 6

J3 8

J4 5
Assume, sequence A - J1-J2-J3-J4
Job Wi ti Ci
J1 0 7 7
MFTsq-A
J2 7 6 13 67/4 = 16.75
J3 13 8 21
J4 21 5 26
41 + 26 67 sequence C –Shortest processing Time(SPT)
sequence B (random) - J2-J3-J1-J4 J4-J2-J1-J3

Job Wi ti Ci Job Wi ti Ci
J2 0 6 6 J4 0 5 5
J3 6 8 14 J2 5 6 11 MFTsq-C
J1 14 7 21 J1 11 7 18 60/4 = 15
J4 21 5 26 J3 18 8 26
41 + 26 67 34 + 26 60

MFTsq-B For single m/c case – SPT guarantees Min. MFT.


67/4 = 16.75
Practice Problem
Consider the following single machine scheduling problem. Find
the optimal sequence which will minimize the mean flow time &
also obtain the minimum mean flow time

Job Processing Time


(hrs)
1 15
2 4
3 5
4 14
5 8
Jobs Processing Time (PT) Flow Time (FT)
1 15 15
2 4 19 Avg. FT
142/5 = 28.4
3 5 24
4 14 38
5 8 46
142

Jobs PT FT SPT – Shortest PT


Jobs PT FT
Avg. FT = 21.4
2 4 4 1 15 15
3 5 9 4 14 29
5 8 17 5 8 37
4 14 31 3 5 42
1 15 46 LPT – Longest PT 2 4 46
107 Avg. FT = 33.8 169
Earliest Due Date (EDD) rule to minimize Max. Lateness

• Maximum job lateness (Lmax) and Maximum job tardiness (Tmax) are
minimized by EDD sequencing.
• In single machine scheduling problem, sequencing the jobs in increasing
order of due date is known as Earliest Due Date rule.

Problem

Consider the following single machine scheduling problem. Find


sequence which minimize (Max. Lateness Lmax), also determine Lmax
with Optimal Sequence

Job J 1 2 3 4 5 6
PT (tj) 10 8 8 7 12 15
DD (dj) 15 10 12 11 18 25
EDD sequence
Lateness = Completion time – Due Date

Arrange the jobs in increasing


order of due dates
Non-EDD sequence

Job PT CT DD Lateness
Job PT CT DD Lateness
2 8 8 10 -2
1 10 10 15 -5
4 7 15 11 4
2 8 18 10 8
3 8 23 12 11
3 8 26 12 14
1 10 33 15 18
4 7 33 11 22
5 12 45 18 27
5 12 45 18 27
6 15 60 25 35 (Max.)
6 15 60 25 35 (Max.)

Lmax of any other non-EDD sequence will not be less than 35


Earliest Due Date (EDD) rule to minimize Max. Lateness

Practice Problem
Consider the following single machine scheduling problem. Determine
the sequence which will minimize the Maximum Lateness (Lmax). Also,
determine the Lmax with respect to the Optimal Sequence

Job J 1 2 3 4 5 6
PT (tj) 8 9 7 8 10 14
DD (dj) 14 12 10 15 20 18
Lateness = Completion time – Due Date

EDD sequence Arrange the jobs in increasing


order of due dates

Job PT CT DD Lateness
3 7 7 10 -3
2 9 16 12 4
1 8 24 14 10
4 8 32 15 17
6 14 46 18 28
5 10 56 20 36 (max)

Lmax of any other non-EDD sequence will not be less than 36


Flow shop scheduling problem
Find out Make Span
Job MC 1 MC 2

1 5 4
2 3 1
3 6 2
4 7 8
Gnatt Chart
Job Sequence: 2-1-4-3 Make span: 25 Idle times on MC 2: 0-3, 4-8, 12-15

MC 1 J2 J1 J4 J3

MC 2 Idle J2 Idle J1 Idle J4 J3


0 3 4 8 12 15 21 23 25
Time
Job Sequence: 3-4-1-2 Make span: 26 Idle times on MC 2: 0-6, 8-13

MC 1 J3 J4 J1 J2

MC 2 Idle J3 Idle J4 J1 J2
0 6 8 13 18 21 25 26
Time
n jobs, Two machines
Johnson’s Algorithm

Step 1. Find the minimum among various ti1 and ti2 .

Step 2a. If the minimum processing time requires machine 1, place the
associated job in the first available position in sequence. Go to step 3.

Step 2b. If the minimum processing time requires machine 2, place the
associated job in the last available position in sequence. Go to step 3.

Step 3. Remove the assigned job from consideration and return to step 1
until all positions in sequence are filled. (Ties may be broken randomly.)
Problem
Consider 2 m/c and 6 job flow shop scheduling problem. Using Johnson’s
algorithm, Obtain optimal sequence which will minimize the makespan
Job 1 2 3 4 5 6
MC 1 5 2 13 10 8 12
MC 2 4 3 14 1 9 11
Solution
Stage Unsheduled Jobs Minimum Assignment Partial
tik Sequence
1 1, 2, 3, 4, 5, 6 t42 4 = [6] XXXXX 4
2 1, 2, 3, 5, 6 t21 2 = [1] 2 XXXX 4
3 1, 3, 5, 6 t12 1 = [5] 2X X X1 4
4 3, 5, 6 t51 5 = [2] 2 5X X1 4
5 3, 6 t62 6 = [4] 2 5X 6 1 4
6 3 t31 3 = [3] 2 5 3 6 1 4

The optimal sequence is 2–5–3–6–1–4

Time-in on MC 2 = Max (MC1 Time-out of current job, MC 2 time-out


of previous job)
Sequence

Jobs Processing Time Idle time


Machine 1 Machine 2 on
Machine 2
Time in Time out Time in Time out
2 0 2 2 5 2
5 2 10 10 19 5
3 10 23 23 37 4
6 23 35 37 48 0
1 35 40 48 52 0
4 40 50 52 53 0

Make span
Johnson's Algorithm

Practice Problem
Consider 2 m/c and 8 job flow shop scheduling problem. Obtain optimal sequence
which will minimize the makespan

Job 1 2 3 4 5 6 7 8
MC 1 (Pj1) 5 2 1 7 6 3 7 5
MC 2 (Pj2) 2 6 2 5 6 7 2 1
Sequence 1

Jobs Processing Time Idle time


Machine 1 Machine 2 on
Machine 2
Time in Time out Time in Time out
3 0 1 1 3 1
2 1 3 3 9 0
6 3 6 9 16 0
5 6 12 16 22 0
4 12 19 22 27 0
1 19 26 27 29 0
7 26 31 31 33 2
8 31 36 36 37 3

Similar procedure for sequence 2, Make span


3-2-6-5-4-7-1-8 Make span 37
Extension of Johnson’s Rule (n jobs, 3 M/C)
Consider 3 M/C and n jobs flow shop scheduling problem
Job Processing Time
MC 1 MC 2 MC 3
1 t11 t12 t13
2 t21 t22 t23
3 t31 t32 t33
. . . .
. . . .
i ti1 ti2 ti3
. . . .
. . . .
n tn1 tn2 tn3

If any one of the below conditions is satisfied then, we can extend the Johnson’s
algorithm
1) if min ti1 ≥ max ti2
2) if min ti3 ≥ max ti2
Extension of Johnson’s Rule
• Create hypothetical problem with two m/c and n jobs as shown in table.
• Objective is to obtain optimal sequence, then makespan is to be determined for
optimal sequence

Job Processing Time


MC A MC B
1 t11 + t12 t12 + t13
2 t21 + t22 t22 + t23
3 t31 + t32 t32 + t33
. . .
. . .
i ti1 + ti2 ti2 + ti3
. . .
. . .
n tn1 + tn2 tn2 + tn3
Johnson’s Algorithm for 3 M/Cs
Consider the following 3 M/Cs and 4 jobs flow shop problem Check whether
Johnson’s rule can be extended to this problem. If so, what is the optimal schedule
and the corresponding makespan?
Job 1 2 3 4
MC 1 (PT) 3 12 8 12
MC 2 (PT) 8 9 6 10
MC 3 (PT) 10 12 13 16
If any one of the below conditions is satisfied then, we can extend the Johnson’s
algorithm
1) if min ti1 ≥ max ti2
2) if min ti3 ≥ max ti2
Condition 2 is satisfied, Job MC 1 + MC2 MC2 + MC
3
1 11 18
2 21 21
3 14 19
4 22 26
Same procedure of 2 M/C problem is applied here to find optimal sequence
Optimal sequence: Sequence 1: 1-3-4-2
Sequence 2: 1-3-2-4
Job Processing Time (in hour) Idle Idle
Machine 1 Machine 2 Machine 3 Time Time
MC 2 MC 3
in out in out in out
1 0 3 3 11 11 21 3 11
3 3 11 11 17 21 34 0 0
4 11 23 23 33 34 50 6 0
2 23 35 35 44 50 62 2 0

Make span

Similar procedure for sequence 2,


CDS Heuristic (n jobs m machines)
• Difficult to get optimal solution in finite time for large size problems.
• Use efficient heuristics for large size problems.
• CDS (Campbell, Dudek and Smith) heuristic is one such heuristic used for flow
shop scheduling.
Stage 1, t1j1 = tj1 and t1j2 = tjm
Johnson’s rule is applied to first and ‘m’th operations and intermediate
operations are ignored.
Stage 2, t2j1 = tj1 + tj2, t2j2 = tjm + tj,m-1
Johnson’s rule is applied to the sum of the first two and the last two
operation processing times. In general at stage i,
𝒊 𝒊
tij1 = 𝒌=𝟏 𝒕𝒋𝒌 and tij2 = 𝒌=𝟏 tj,m−k+1
For each stage i (i=1,2,…, m-1), the job order obtained is used to calculate
a makespan for the original problem. After m-1, stages, the best makespan among
the m-1 schedules is identified. (some of the m-1 sequences may be identical).
Find the makespan using the CDS heuristic for the following flow shop problem
Job j tj1 tj2 tj3 tj4
1 4 3 7 8 Stage 1 Optimal Sequence: 3-2-1-4
2 3 7 2 5
job MC 1 (tj1) MC 4 (tj4)
3 1 2 4 7
1 4 8
4 3 4 3 2
2 3 5
3 1 7
4 3 2

Job Processing Time (in hour)


Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 1 1 3 3 7 7 14
2 1 4 4 11 11 13 14 19
1 4 8 11 14 14 21 21 29
4 8 11 14 18 21 24 29 31
Stage 2 job MC 1 + MC2 MC 3 + MC 4
1 7 15
2 10 7
3 3 11
4 7 5

Optimal Sequence: 3-1-2-4


Job Processing Time (in hour)
Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 1 1 3 3 7 7 14
1 1 5 5 8 8 15 15 23
2 5 8 8 15 15 17 23 28
4 8 11 15 19 19 22 28 30
job MC 1 + MC 2 + MC 3 MC 2+ MC 3+ MC 4
Stage 3 1 14 18
Optimal Sequence: 2 12 14
3-2-1-4 3 7 13
4 10 9

Job Processing Time (in hour)


Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 1 1 3 3 7 7 14
2 1 4 4 11 11 13 14 19
1 4 8 11 14 14 21 21 29
4 8 11 14 18 21 24 29 31
Stage Sequence Makespan
Summary best
1 3-2-1-4 31 sequence
2 3-1-2-4 30
3 3-2-1-4 31
Practice Problem
Find the makespan using the CDS heuristic for the following flow shop problem

Job j tj1 tj2 tj3 tj4


1 5 4 8 9
2 4 8 3 6
3 2 3 5 8
4 4 5 4 3
Stage 1 job MC 1 (tj1) MC 4 (tj4)
1 5 9
2 4 6
3 2 8
4 4 3

Optimal Sequence: 3-2-1-4


Job Processing Time (in hour)
Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 2 2 5 5 10 10 18
2 2 6 6 14 14 17 18 24
1 6 11 14 18 18 26 26 35
4 11 15 18 23 26 30 35 38
Stage 2 job MC 1 + MC2 MC 3 + MC 4
1 9 17
2 12 9
3 5 13
4 9 7

Optimal Sequence: 3-1-2-4


Job Processing Time (in hour)
Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 2 2 5 5 10 10 18
1 2 7 7 11 11 19 19 28
2 7 11 11 19 19 22 28 34
4 11 15 19 24 24 28 34 37
Stage 3 job MC 1 + MC 2 + MC 3 MC 2+ MC 3+ MC 4
1 17 21
Optimal Sequence: 2 15 17
3-2-1-4 3 10 16
4 13 12

Job Processing Time (in hour)


Machine 1 Machine 2 Machine 3 Machine 4
in out in out in out in out
3 0 2 2 5 5 10 10 18
2 2 6 6 14 14 17 18 24
1 6 11 14 18 18 26 26 35
4 11 15 18 23 26 30 35 38

Summary Stage Sequence Makespan best


1 3-2-1-4 38 sequence
2 3-1-2-4 37
3 3-2-1-4 38

S-ar putea să vă placă și