CHAPTER ONE

INTRODUCTION

1.1 Introduction

1
1.2 Sequencing and Scheduling

Scheduling is the task of determining when each operation is to start and finish. Since
each operation is in possible competition with other operations for scarce resources of
time and capacity, the job of scheduling is neither simple nor easy.Scheduling is the
allocation of resources over time to perform a collection of tasks. This rather general
definition of the term does convey two different meanings. First scheduling is a
decision-making function. Second scheduling is a body of theory as it is a collection
of principles, models, techniques and logical conclusions that provide insight into the
scheduling function.It may be worth mentioning here to distinguish between terms
"Scheduling" and "Sequencing" which as a matter of fact, are both associated with the
job shop process. Scheduling is defined as assigning each operation of each job a start
time and a completion time on a time scale of machine within the precedence
relations. However, under certainty, a schedule cannot be determined because the
arrival time of the job, operation processing time, as well as other attributes are not
fully known. On the other hand, sequencing means that for each machine in the shop,
one has to establish the order in which the jobs waiting in the queue in front of that
particular machine have to be processed.The problem that motivated this study is as
follows: suppose there are a number of jobs to be performed. Each job consists of a
given sequence of operations which needs to be performed using a number of
machines. All operations for each job must be performed in the order given by the
sequence. Each operation demands the use of a particular machine for a given time.
Each machine can process only one operation at a time. Therefore, given a cost
function by which each sequence can be evaluated, the order of operations on each
machine thatminimizesthe cost function needs to be found[1]

1.2.1 Single-machine sequencing

In this type of scheduling jobs are first arranged in a particular order in a given
machine. Our objective of minimizing the make span is best fulfilled when the
arrangement of jobs is done on the basis of in process time, that means jobs with less
in process time are put forward than those having higher in process times. [3]

1.2.2Flow shop scheduling

2
The basic principle of this type of scheduling is that each machine takes the jobs in a
particular sequence, which means each job has to get processed in each and every
machine available in the flow shop environment. Also, in each machine the in-process
time for each and every job is distinct. Therefore, while evaluating a flow shop
problem, we carry out different possible sequences of carrying out the job, and then
the best among those is chosen.[3]

1.2.3 Job shop scheduling

The basic difference between job shop scheduling and flow shop scheduling is that
unlike flow shop, here it is not compulsory for all the jobs to get processed in each
and every machine available, that means each distinct job may be processed in any
distinct number of machines as requirement demands. The sequence of jobs to be
processed is also distinct in each machine. Based on the requirement of the problem,
here also we choose the best sequence out of a number of possible sequences. [3]

1.3 Scheduling Environments

Sequencing and scheduling environments are classified according to four types of

information: the jobs and operations to be processed; the number and types of
machines that comprise the shop; the disciplines that restrict the manner in which
assignment can be made, and the criteria by which a schedule will be evaluated. The
sequencing and scheduling environments are as follows:

1. Single machine shop: one machine and n jobs to be processed.

2. Flow shop: there are m machines in series and jobs can be processed in one of
the following ways:
a) Permutational: jobs are processed by a series of m machines in exactly the
same order, or
b) Non-permutational: jobs are processed by a series of m machines not in the
same order.
3. Job shop: each job has its flow pattern and a subset of these jobs can visit each
machine twice or more often. Multiple entries and exits [1]

3
1.4 Role of Fuzziness in Scheduling

In the literature dealing with job sequencing problems, the processing time is assumed
to be known exactly. However, in practical situations, this is seldom the case. In
general, the processing time can only be estimated as within a certain interval. The
particular characteristics of this processing time interval can be exactly represented by
a fuzzy number. Thus, fuzzy set theory is ideally suited for solving job sequencing
problems. One of the problems in using fuzzy set theory to solve job sequencing
problems is that fuzzy numbers only form partial order and thus comparison of fuzzy
numbers to obtain a total or linear order can be a problem. [2]

1.5 Objectives of The Study

This thesis is purely dedicated to the scheduling of jobs in flow shop and job shop
environment, based on the following objectives: -

a) To obtain a minimized make span value by different methods in job shop and
flow shop scheduling problems.
b) To obtain an optimal sequence for every machine and job. This would indicate
the sequence of processing the jobs which would lead to achieve the desired
result.
c) To measure the performance of selected heuristics by comparing the results of
each other in a flow shop environment. [3]

4
 References

1. Alharkan M. Ibrahim, Algorithms for Sequencing and Scheduling, King Saud

University, College of Engineering, Industrial Engineering Department.
2. McCAHON. S. C, LEE. S. E, 1990, "Job Sequencing With Fuzzy Processing
Times", Computers Math. Applic, Vol. 19, No.7, pp. 31-41, 1990.
3. Manoj Kumar Das, Selected Heuristic Algorithms for Solving Job Shop and
Flow Shop Scheduling Problems, Rourkela – 76900, National Institute Of
Technology, Department of Mechanical Engineering.

5
 CHAPTERTWO

REVIEW OF SOME PREVIOUS RELATED WORKS

Saharanpur at el,studied the Fuzzy numbers were ideally suited to represent

uncertainty. In the present paper hehave generalized & modified deterministic
approach to accept LR - type fuzzy processing time, the sequence performance
measurements of make span and job mean flow time are fuzzy in nature. Earlier
“McMahon & Lee proposed an algorithm for managing uncertain scheduling.
However they have used trapezoidal fuzzy number to represent the processing time.
After that “T.P. Hong & T.N. Chuang proposed a triangular fuzzy Johnson algorithm
for n x 2 job shop scheduling problem they have used half inverse operator to
compare fuzzy number as well as to find triangular longer time procedure. In the
present work he have applied GRV (Generalized Ranking Value) technique for the
generalized LR - type fuzzy number. “McCahon &Lee “TP Hong & TN Chuang
algorithm can be special case for our proposed algorithm.

In the present work LR-type fuzzy number have been used with Johnson algorithm &
CDS algorithm to schedule n-jobs on m-machine with uncertain completion time. The
new generalized fuzzy Johnson algorithm yields scheduling results with a
membership function for the final completion time. These results can help managers
gain a broader overall view of scheduling. In future they will try to apply other
characteristics of fuzzy sets to the scheduling fields

TZUNG-PEI HONG* AND TZUNG-NAN CHUANG at el, In the past, we have

demonstrated how fuzzy concepts can easily be used in the Johnson algorithm to
manage uncertain scheduling on two-machine flow shops. This paper extends
application to fuzzy flow shops with more than two machines. A new fuzzy heuristic
flow-shop scheduling algorithm (the fuzzy Palmer algorithm) is then designed since
optimal solutions seem unnecessary for uncertain environments. Also, the
conventional Palmer algorithm is presented as a special case of the fuzzy Palmer
algorithm with special assigned membership functions..

In this paper, we have proposed the fuzzy Palmer algorithm for scheduling uncertain
jobs in a flow shop with more than two machines. The fuzzy Palmer algorithm can
yield a scheduling result with a membership function for the final completion time.
The results can then help managers gain a broader overall view of scheduling. Also,
the conventional Palmer algorithm has been shown to be a special case of the fuzzy
Palmer algorithm with special assigned membership functions. The fuzzy Palmer
algorithm is then a feasible solution for both deterministic and uncertain flow shops
with more than two machines. In the future, we will try to apply other characteristics
of fuzzy sets to the scheduling field

6
Hamirpur at el, Scheduling ‘n’ jobs on ‘m’ machines in a flow shop is NP- hard
problem and places itself at prominent place in the area of production scheduling. The
essence of any scheduling algorithm is to minimize the makespan in a flowshop
environment. In this paper an attempt has been made to develop a heuristic algorithm,
based on the reduced weightage of machines at each stage to generate different
combination of ‘m-1’ sequences. The proposed heuristic has been tested on several
benchmark problems of Taillard (1993) [Taillard, E. (1993). Benchmarks for basic
scheduling problems. European Journal of Operational Research, 64, 278-285.]. The
performance of the proposed heuristic is compared with three well-known heuristics,
namely Palmer’s heuristic, Campbell’s CDS heuristic, and Dannenbring’s rapid
access heuristic. Results are evaluated with the best-known upper-bound solutions and
found better than the above three

In this paper, we have presented a heuristic for the general flow shop scheduling to
minimize the makespan. The proposed method was based on the principle that
weightage of the machines at each stage was reduced to obtain different combination
of sequences. The sequence with minimum makespan is selected as the best sequence.
The heuristic was tested using various benchmark problems taken from Taillard. The
percentage gaps with best-known upper bound value were also tabulated. The
computational results indicate that the proposed heuristic significantly performed
better than the heuristics of CDS, Palmer and RA. Also, it can been seen that as the
number of jobs increases, proposed heuristic provides good quality results. Therefore,
it is the main reason to recommend this heuristic mainly for large size problems.
Future scope of this research provides the extensive use of proposed heuristics for
researchers to develop hybrid heuristics / metaheuristics for solving flow shop
scheduling problems and use of this algorithm for the generation of initial solutions
because of the superiority over existing heuristic algorithms.

C. S. MCCAHON 1 and E. S. LEE 2 at el, In practice, processing times can be more

accurately represented as intervals with the most probable completion time
somewhere near the middle of the interval. A fuzzy number which is essentially a
generalized interval can represent this processing time interval exactly and naturally.
In this work, triangular and trapezoidal fuzzy numbers are used to represent those
vague job processing times in job shop production systems. The job sequencing
algorithms of Johnson and Ignall and Schrage are modified to accept fuzzy job
processing times. Fuzzy makespans and fuzzy mean flow times are then calculated for
greater decision-making information. Numerous examples are used to illustrate the
approach..

The job sequencing algorithms of Johnson and Ignall and Schrage were modified to
accept two types of fuzzy job processing times: triangular and trapezoidal. The
resultant job sequences were non-fuzzy, but the performance criteria of makespan and
mean flow time were fuzzy using these modified algorithms. The fuzzy performance

7
criteria could then be interpreted using possibility theory and/or fuzzy integrals. In
some cases, if the fuzzy processing times can be represented using triangular fuzzy
numbers, the optimal sequence and means of the fuzzy makespan and fuzzy mean
flow time can be determined by using a deterministic approximation. This
approximation is made by using the modes (b values) of the triangular processing,
times as deterministic with the non-modified (or original) sequencing algorithm. The
problem with using this approximation is that the fuzzy makespan and mean flow
time must be calculated anyway if further sensitivity analysis using possibility theory
and/or fuzzy integrals is desired. If trapezoidal fuzzy numbers are used, then the
fuzzified method must be used. In summary, by keeping the fuzziness throughout the
analysis procedure, the decision maker keeps intact information useful for subsequent
sensitivity analysis of the performance criteria. If the input job processing times are
truly fuzzy, then they should be modeled as fuzzy to obtain fuzzy results.
Approximations can be used, but only in special cases, and without the
comprehensive results obtained from the fuzzy procedure.

N.Selvamalar, V.Vinoba at el, In this paper, the three fuzzy heuristics namely fuzzy
Campbell-Dudek-Smith algorithm(FCDS), fuzzy NawazEnscore-Ham(FNEH)
algorithm and fuzzy Rapid Access(FRA) algorithm has been subjected to a
comparative study to evaluate and assess the efficiency of the fuzzy models in various
problems of random size.A sample data is chosen and 14 problems are generated with
different job and machine size.The makespan is evaluated,compared and depicted
using ORIGIN software.The worst case solution of every problem is compared with
the heuristic schedule obtained in order to show the improvement in the solution when
the above heuristics are applied.It is noticed that the FNEH algorithm stand tall above
the remaining procedures adopted in this work except a few instances where the FRA
procedure excels.

When the machine size is less as 4 machines or 5 machines,all the procedures behave
similarly in producing the results. Therefore it is better to consider the results from
RA algorithm, since it is very fast in giving the good solution to the problem,.But
when the machine size increases, the NEH algorithm gives the better result compared
to CDS and RA algorithms. Whatsoever be the problem size,one could see that NEH
algorithm produces better results compared to CDS and RA algorithms in the case of
5 jobs problems.When the machine size is less, the difference in makespan obtained
by the three methods seems negligible but when the machine size goes above 5, we
could see the advantage of using the NEH algorithm. There is a gradual growth in the
makespan due to the three heuristics and it is evidently seen when the machine size
starts increasing .The objective of minimizing the makespan is very well satisfied by
the NEH procedure. The fuzzy NEH algorithm outshines the other algorithms as it
shows great difference in the makespan.In this case also fuzzy NEH proves to be best.
Therefore on have to conclude that the FNEH algorithm stands tall among all the
other methods irrespective of its computational complexity.The future research should

8
be in the direction of decreasing its complexity.But when shorter computational time
is required ,it is better to go with FRA procedure because it has come to light that
FRA is also providing equally good results. Considering the schedule which is
constructed by arranging the jobs in the descending order of the sum of their
execution times as the worst case ,the heuristic solutions are compared and it is seen
that an improvement is done upto 17%.If we find the lower bound for the
makespan ,we would be able to get the goodness level of the heuristic solutions as
well.

Salwani Abdullah, Majid Abdolrazzagh-Nezhad at el, Fuzzy job-shop scheduling

problems (Fuzzy JSSPs) are a class of combinational optimization problems known as
non-deterministic polynomial-hard problems. In recent decades, a number of
researchers have expanded the theoretical models of Fuzzy JSSPs and introduced
algorithms to solve them. This paper reviews the classification of Fuzzy JSSPs,
constraints and objectives investigated in Fuzzy JSSPs, and the methodologies applied
in solving Fuzzy JSSPs. The paper centers on reviewing meta-heuristic algorithms as
state-of-the-art algorithms proposed for Fuzzy JSSPs. These algorithms are analyzed
in three steps, namely, pre-processing, initialization procedures, and improvement
algorithms. Finally, possible suggestions for future studies are obtained from this
survey.

E.O. Oyetunji at el, : In this Study, we discussed 29 different scheduling criteria. We

developed mathematical expressions for all the criteria considered. Each of the criteria
was expressed as a function of either the completion time of job or the given
parameters. This will assist researchers to easily compute the value of any of the
scheduling criteria considered in this study.

About 29 distinct scheduling objectives have been discussed and the expressions for
computing their values have been formulated from the first principle. The objectives
are all expressed in terms of the completion time of jobs and the given parameters
thereby simplifying their use.

2http://growingscience.com/ijiec/VOL1/IJIEC_2010_7.pdf