Scheduling of Operations

What is scheduling?
Last stage of planning before production occurs Planning and control issues pertaining to short-term (ranging from a day/shift to utmost a week or two) In short-term several questions need to be answered.

How can the jobs be assigned to various work centers? Within each work center, how can the jobs be rank ordered? How can other resources such as skilled workers and special gadgets be assigned to the operating system?

A collective framework to address these questions is referred to as scheduling of operations.

Objectives in Scheduling
Meet customer due dates Minimize job lateness Minimize completion time Minimize overtime Maximize machine or labor utilization Minimize idle time Minimize work-in-process inventory

Terminologies

Loading: Resources in an operating system are assigned with adequate number of jobs during the planning horizon. Ensures that the resources in an operating system are utilized in the best possible way. Routing: The technological order in which resources available in a shop are used by the job for processing. Scheduling & Sequencing: Process of rank ordering the jobs according to some chosen performance measures for each resource. Dispatching: Administrative process of authorizing processing of jobs by resources in the operating system, as identified by the scheduling system, in the form of shop order or information communicated electronically. Expediting: Administrative process of reviewing the progress of the job in the operating system and ensuring that it travels through various stages of the process as indicated in routing, without harming the performance measures. If there are deviations and delays, it calls for pushing the job forward through the system.

Loading
Process of assigning work to limited resources Perform work on most efficient resources Use assignment method of linear programming to determine allocation

Assignment Method
1. Perform row reductions
1.

subtract minimum value in each row from all other row values subtract minimum value in each column from all other column values use minimum number of horizontal and vertical lines

2. Perform column reductions

1.

3. Cross out all zeros in matrix

1.

4. If number of lines equals number of rows in matrix then optimum solution has been found. Make assignments where zeros appear 5. Else modify matrix
1. 2.

3.

subtract minimum uncrossed value from all uncrossed values add it to all cells where two lines intersect other values in matrix remain unchanged

6. Repeat steps 3 through 5 until optimum solution is reached

Assignment Method: Example

Initial Matrix Bryan Kari Noah Chris Row reduction 5 4 2 5 0 0 1 1 1 2 0 0 5 4 1 6 PROJECT 3 4 6 10 4 6 5 6 4 10 Cover all zeros 3 2 0 3 0 0 1 1 1 2 0 0 4 3 0 5

1 10 6 7 9

2 5 2 6 5

Column reduction 3 2 0 3 0 0 1 1 1 2 0 0 4 3 0 5

Number lines number of rows so modify matrix

Assignment Method: Example (cont.)

Modify matrix Cover all zeros

1 0 0 1

0 0 3 1

1 2 2 0

2 1 0 3

1 0 0 1

0 0 3 1

1 2 2 0

2 1 0 3

Number of lines = number of rows so at optimal solution 1 1 0 0 1

Bryan Kari Noah Chris

PROJECT 2 3 0 1 0 2 3 2 1 0

4 2 1 0 3

Bryan Kari Noah Chris

1 10 6 7 9

PROJECT 2 3 4 5 6 10 2 4 6 6 5 6 5 4 10

Project Cost = 5 + 6 + 6 + 4 = 21

Scheduling Context
The following parameters adequately describe the scheduling problem.

Number of machines/resources Number of jobs Shop configuration: the manner in which machines are organized on the shop floor and the flow pattern of the jobs utilizing the machines.

Flow shop resources are organized one after other in the order the jobs are processes. If there are n jobs, there are n! ways in which one can draw up alternative schedule. Job shop machines are not organized in processing order, rather similar type of resources is grouped together. If there are n jobs and m machines, there are (n!)m ways in which one can draw up alternative schedule.

Scheduling Rules
SPT - shortest processing time LPT - longest processing time EDD - Earliest Due Date CR - critical ratio

Considers both processing time and due date information. CR estimates the criticality of the job as CR = Remaining time / Remaining work = (Due date Current date) / Remaining processing time Smaller the value of CR indicates the job is more critical. If CR < 1, indicates the available time is not sufficient and the job is already running late and if CR > 1, indicates some slack is available for the job.

FCFS - first-come, first-served LCFS - last come, first served RAN Random order CUSTPR - highest customer priority SETUP - similar required setups

EXAMPLE
Current time = 0 Order of arrival 1 2 3 4

Job No 1 2 3 4

Processing time 12 9 22 11

Due by 23 24 30 20

SPT 3 1 4 2

LPT 2 4 1 3

EDD 2 3 4 1 RAN

FCFS 1 2 3 4

CR 1.917 2.667

Order 3 4

1.364
1.818

1
2

Random No

0.8403
0.9930 0.6439 0.4418

3
4 2 1

Performance criterion
Completion-based measures

Assessment of the processing time is done. Frequently used measures:

Flow Time the elapsed time between releasing a job into the shop and the time of completion of processing of the job. It is equal to the sum of processing time of all operations of the job and other non-productive time including waiting and moving time.
Release time of the job be Ri, Completion time of the job be Ci, then

Flow time Fi = Ri Ci

Make span the time taken to complete all the jobs released into the shop for processing. It is the completion time of the last job in the shop.
Make span (maximum completion time) Cmax = max { Ci }

Performance criterion (cont)

Due date-based measures

Efficacy of the scheduling rule is evaluated with reference to the due date of the jobs. Measures:

Lateness: The difference between completion time and due date.

If the due date is denoted as Di, then lateness Li = Ci Di

Tardiness: Captures information of only those jobs that are late. Negative values of lateness are considered to be 0.
Tardiness Ti = max(0, Li)

Number of tardy jobs nT

Utilization-based measures

Based on actual inventory build up in the shop or utilization of the shop resources. Measures include the mean number of jobs waiting, mean number of finished jobs, Simulation models are typically used such scheduling problems.

Example
PROCESSING TIME DUE DATE

JOB

A B C D E

5 10 2 8 6

10 15 5 12 8

Sequencing Rules: FCFS

FCFS SEQUENCE

START TIME

PROCESSING COMPLETION DUE TIME TIME DATE

TARDINESS

A B C D E

0 5 15 17 25

5 10 2 8 6

5 15 17 25 31

10 15 5 12 8

0 0 12 13 23

Sequencing Rules: EDD

EDD SEQUENCE

START TIME

PROCESSING COMPLETION DUE TIME TIME DATE

TARDINESS

C E A D B

0 2 8 13 21

2 6 5 8 10

2 8 13 21 31

5 8 10 12 15

0 0 3 9 16

Sequencing Rules: SLACK

A(10-0) 5 = 5 B(15-0) - 10 = 5 C(5-0) 2 = 3 D(12-0) 8 = 4 E(8-0) 6 = 2

SLACK SEQUENCE

START TIME

PROCESSING COMPLETION DUE TIME TIME DATE

TARDINESS

E C D A B

0 6 8 16 21

6 2 8 5 10

6 8 16 21 31

8 5 12 10 15

0 3 4 11 16

Sequencing Rules: CR

A(10)/5 = 2.00 B(15)/10 = 1.50 C (5)/2 = 2.50 D(12)/8 = 1.50 E (8)/6 = 1.33
TARDINESS

CR SEQUENCE

START TIME

PROCESSING COMPLETION DUE TIME TIME DATE

E D B A C

0 6 14 24 29

6 8 10 5 2

6 14 24 29 31

8 12 15 10 5

0 2 9 19 26

Sequencing Rules: SPT

SPT SEQUENCE

START TIME

PROCESSING COMPLETION DUE TIME TIME DATE

TARDINESS

C A E D B

0 2 7 13 21

2 5 6 8 10

2 7 13 21 31

5 10 8 12 15

0 0 5 9 16

Sequencing Rules: Summary

AVERAGE COMPLETION TIME AVERAGE TARDINESS NO. OF JOBS TARDY MAXIMUM TARDINESS

RULE

FCFS DDATE SLACK CR SPT

18.60 15.00 16.40 20.80 14.80

9.6 5.6 6.8 11.2 6.0

3 3 4 4 3

23 16 16 26 16

A manufacturer is in the process of assessing the usefulness of SPT and EDD rules for the purpose of scheduling four jobs on a machine. The table below gives the related information. Compute relevant performance measures and comment. Job No 1 2 3 4 Process time 4 7 2 8 Due by 6 9 19 17

A flow process with two machines and n number of jobs to be scheduled.

Johnsons Rule
List time required to process each job at each machine. Set up a one-dimensional matrix to represent desired sequence with number of slots equal to number of jobs. Select smallest processing time at either machine. If that time is on machine 1, put the job as near to beginning of sequence as possible. If smallest time occurs on machine 2, put the job as near to the end of the sequence as possible. Remove job from list. Repeat until all slots in matrix are filled and all jobs are sequenced. The resulting sequence of the jobs is the best schedule to minimize the make span of the jobs.

EXAMPLE: Using Johnsons Rule

JOB A B C D E

PROCESS 1 6 11 7 9 5

PROCESS 2 8 6 3 7 10

Johnsons Rule (cont.)

E
E 5 A 11 D 20

D
B

C
C 31 38
Process 1 (sanding)

Idle time E 5 15 A 23 D 30 B 37 C 41
Process 2 (painting)

Completion time = 41 Idle time = 5+1+1+3=10

Guidelines for Selecting a Sequencing Rule

1. SPT most useful when shop is highly congested 2. Use SLACK for periods of normal activity 3. Use EDD when only small tardiness values can be tolerated 4. Use LPT if subcontracting is anticipated 5. Use FCFS when operating at low-capacity levels 6. Do not use SPT to sequence jobs that have to be assembled with other jobs at a later date

Gantt Chart
Job 32B
3 Job 23C 2 Job 11C 1 Ahead of schedule Job 12A On schedule Behind schedule

Facility

1 Key:

6 8 Todays Date

10

11

12

Days

Planned activity
Completed activity