KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY

VOL.I, No.III, JANUARY, 2007.

A FUZZY LINGUISTIC APPROACH OF PREVENTIVE MAINTENANCE

SCHEDULING COST OPTIMISATION

Oke, S.A1., Charles-OwabaO.E. 2

1
Department of Mechanical Engineering, University of Lagos, Nigeria
2
Department of Industrial and Production Engineering, University of Ibadan, Nigeria

Corresponding Author E-Mail: [email protected]

ABSTRACT
The simultaneous scheduling of resource-constrained maintenance and operations is an
emerging theory that is attracting the attention of both researchers and practitioners. The
purpose of this paper is to capture the uncertainty in the development of a model that
schedules both preventive maintenance and operational activities. Fuzzy logic is employed to
transform the human expertise into IF-THEN rules. The approach has the advantage of
revealing semantic uncertainty with the associated non-specifying measures. The
methodology applied tracks the error values in terms of results in linguistic variable. The
result obtained indicates the feasibility of tracking the uncertain measures in the model
discussed. Thus, the study may be applicable to both production system and transportation
organizations that are engaged in both maintenance and operational activities. The research
has serious implication in terms of the ability to monitor the imprecision that were introduced
in the previous models. This obviously provides a more reliable framework for researchers
and practitioners interested in maintenance scheduling operation. The paper is new in that, it
demonstrates the application of fuzzy logic in a form that was never documented.

INTRODUCTION
Many organization worldwide are shifting their focus from the traditional way of running the
business to a modern and pro-active approach that guarantees the continued existence of the
business. Traditionally, businesses are managed with the sole aim of making profit. This
viewpoint is fading away in view of the emerging business improvement concepts that has
dominated organizational practices worldwide. Such concepts include business process re-
engineering, total quality management, etc. As such the customer is at the central focus of
the company’s objective with the aim of satisfying and exceeding customers' requirements.

This condition has created intense competition among rival organizations. Therefore, for
businesses to stay afloat, customer satisfaction should be the central focus. While satisfying
the customer, the profit maximization objective of company would automatically be achieved
since an increased customer patronage is envisaged. In an effort to continue in business,
company managers are re-engineering their processes for efficiency and effectiveness. The
maintenance culture is one of the subsystems of the organization that has received an
increased attention in recent times. Fortunately the viewpoint of the maintenance function as
a “bottomless pit of expenses” is fading away. Maintenance is now viewed as a “value
adding activity” to the organization. In other to improve efficiency of operations, proper
scheduling of facilities should be done. In scheduling facilities, optimal approaches are
sought for such that minimum cost is expended on activities and maximum results obtained.
Until recently most approaches utilized in maintenance scheduling are near optimal (see
Anily et al., 1998, 1999). This is largely due to the non-polynomial problem that usually
exists in problem formulation.

1
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

Fortunately, Charles Owaba (2002) has recently proposed a breakthrough methodology that
uses the basic principles of Gantt Charting. Since the development of the original model
numerous efforts have been made by authors in extending its framework. Such extensions
have incorporated a reviewed cost dimensions, and alternative consideration without the use
of Gantt charting principles. Unfortunately, none of the articles have considered the
uncertain nature of the problem semantic uncertainty in the model framework is an important
concern that is worthy of analytical investigation (Oke, 2004a,b). This important gap in the
maintenance scheduling literature is addressed in the current work.

In particular, the fuzzy logic concept is applied to the simultaneous scheduling of resource-
constrained maintenance and operations. The linguistic variables are used to substitute for the
key-parameters and variable of the Gantt charting model. The model that evolves from the
current research is of immense benefits both to the academic community and the industry.
With the model, the real value of measure could be easily captured while the decisions based
on the modeling information will be more reliable than the former model that does not
contain fuzzy logic concepts.

Literature review: The maintenance scheduling literature has a variety of applications (Zurn
and Quintana, 1975; Wang et al., 2002; Walker et al., 2001a,b). These diverse applications
include aircraft maintenance, process industry, vehicle fleet maintenance, railway track
maintenance, power generation, pavement maintenance, highway maintenance, refinery and
production facilities (Alfares, 1999; Ashayeri, et al., 1996; de Campos and Belhot, 1994;
Gunn and Lee, 1991; Lee, 1991; Lake et al., 2000, 2001; Lake and Ferreira, 2001; Creemers
et al., 1994). An attempt to codify the maintenance scheduling literature portrays a line of
inquiry that has grown in volume and in depth. Despite the immense resources allocated to
the study of maintenance scheduling during the past 30 years, a more intensive study into
some deep structures is still required.

Subsequently, a number of theoreticians proposed models for scheduling maintenance (Zurn

and Quintana, 1977; Stremel and Jenkins, 1981; Stremel, 1981; Ram and Olumolade, 1987;
Duffuaa and Al-Sltan, 1997). Duffuaa and Ben-Daya (1994) proposed a model, which is
similar to that of Hariga in that it focused on non-identical production units (Duffuaa and
Ben-Daya, 1994). The model is developed for the joint overhaul problem by incorporating
the cost of production, and the cost of coordination. The model obtains lower costs of two
examples. Among the deficiencies of these models is the fact that no evidence of machine
dominance concept is considered.

Cheung and Hui (2004) investigated a chemical production site from the perspective of loss
minimization. The paper proposes a multi-period mixed integer linear programming (MILP)
model as an aid to optimize short-term maintenance schedule. In a study by El-Sharkh and
El-keib (2003), the concept of a fuzzy evolutionary programming-based solution
methodology for security-constrained generator maintenance scheduling was explored. The
paper presents the fuzzy model with uncertainties in the load and fuel and maintenance costs.
The technique results are fuzzy optimal cost range that reflects the problem uncertainties. It
solves a decomposed maintenance model of two interrelated subproblems, namely the
maintenance and the security-constrained economic dispatch problem. Dieulle et. al. (2003)
focused on the development of a new probabilistic method based on the semi-regenerative
property of the evolution process. In order to calculate the long-time expected cost per unit
of time, the authors use a recent result generalizing the well-known theorem that expresses

2
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

the cost criterion as equal to the ratio of the expected cost on renewal cycle over the expected
cycle duration. Numerical experiments show that there exists a set of parameters (the critical
threshold and the parameters of maintenance scheduling function) which lead to a minimal
cost.

El-Sharkh and El-Keib (2000) considered an evolutionary programming (EP) based technique
to the unified model of the maintenance scheduling (MS) problem of power generation and
transmission systems. The Hill-climbing technique (HCT) is used in conjunction with the EP
to find a feasible solution in the neighbourhood of the new feasible solutions during the
solution process.

The work by Kobbacy et al. (1997) is concerned with the development of a realistic
preventive maintenance (PM) scheduling model. A heuristic approach for implementing the
semi-parametric proportional hazard model (PHM) to schedule the next preventive
maintenance interval on the basis of the equipment’s full condition history was introduced.
These models are then used within a simulation framework to schedule the next preventive
maintenance interval. The results indicate a higher availability for the recommended
schedule than the availability resulting from applying the optimal PM intervals as suggested
by using the conventional stationary models.

Olorunniwo and Izuchukwu (1991) applied the concept of maintenance improvement factors
to preventive and overhaul maintenance. The authors developed mathematical models that are
used to generate preventive and overhaul maintenance schedules. Examples are provided to
demonstrate the sensitivity of the schedules to model parameters.

Ashayeri et al. (1996) reported on simultaneous planning of preventive maintenance and

production in a process industry environment. The authors developed a mixed integer linear
programming model. The model scheduling production jobs and preventive jobs, while
minimizing costs associated with production back orders, corrective maintenance and
preventive maintenance. The model takes into account the probability of a breakdown given
by the last maintenance period. The formulation of the model is flexible so that it can be
adapted for several production situations. However, the deterministic model discussed does
not afford insight into the nature of deterioration of machines. A common form of machine
deterioration is stochastic but many studies have assumed linear deterioration for ease of
modeling.

Model framework: From the original proposal by Charles-Owaba (2002) the problem is
defined as follows:

Given a set of machines for preventive maintenance and operations in T contiguous periods,
limited periodic maintenance capacity (Aj), limited budgetary allocation (Ba); limited
manpower resources (Mr); duration (Bir) per maintenance visit, arrival period Kij); and the
number of visits per machine (Ni); select the periods for alternative preventive maintenance
and operations under conditions of inflation, such that the total preventive maintenance cost
is minimum (see figure 1).

3
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

criteria parameter constraint variables

preventive
maintenance cost activities machine idle

limited
maintenance
operational
time activities operation

operational
profit

Figure 1: Structural Model of Charles-Owaba’s OGC Model

The following assumptions, adopted from Charles-Owaba (2002) are used in generating
fuzzy logic model.
i. Time for corrective and preventive maintenance is considered.
ii. Cycle of activities occurs in the following order: Operation – maintenance.
iii. Total number operation – maintenance periods are not fixed.
iv. Maintenance periods are known.
v. An arriving machine has its maintenance activities commenced only when resources
are available. Otherwise, it waits.
vi. A machine is either operating or in for maintenance at any moment.
vii. Arrivals occur and maintenance of operation activities commences at the beginning of
a period while completion is at the end of a period.
viii. Arrival periods for planned preventive maintenance are pre-scheduled.
ix. The span of a period maybe one or more seconds, minutes, hours, days, weeks, month
etc.
x. All assumptions are ensured to work together for the common objective of
maximizing profit by cost minimization.

Parameters used in fuzzy logic model

i. Operation cost → (OPC)
ii. Maintenance cost → (MC)
iii. Projected cost→ (PRC)

These parameters constitute the fuzzy logic variables used to generate the fuzzy logic model.
The following error terms of the model are thus defined.

Error terms
PRC – (Σ OPC / MC) = Z, “Zero – error” term (No change in projected cost)
PRC – (Σ OPC / MC) = N, “Negative – error” term (Greater than projected cost)
PRC – (Σ OPC / MC) = P. “Positive – error” term (Less than projected cost)

4
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

Considering the model over length of time, we have:

d{PRC – (Σ OPC / MC)}/dt = Z , “Zero-error-dot” (No change in projected cost overtime)

d{PRC – (Σ OPC / MC)}/dt = N , “Negative-error-dot” (Greater than projected cost overtime)

d{PRC – (Σ OPC / MC)}/dt = P , “Positive-error-dot” (Less than projected cost overtime)

It is reasonable to consider conditions of the model as changing or varying to a large degree

overtime, hence, the following additional terms are generated for more effective control of
the fuzzy logic model.

d{PRC – (Σ OPC/MC)}/dt = RZ , “Zero-error-dot” remains (No change condition remains

overtime)
d{PRC–(ΣOPC/MC)}/dt= CRZ ,“Negative-error-dot”(No change remains continuous overtime)
d{PRC–(ΣOPC/MC)}/dt= >> N  , “Positive-error-dot” (Grreater than projected cost overtime)
d{PRC–(ΣOPC/MC)}/dt= >>> N  , very, very large “Negative-error-dot” (Very, very greater
than projected cost overtime)
d(PRC–{ΣOPC/MC)}/dt = >> P , “very, very large positive -error-dot” (Very, very Less than
projected cost overtime)
d{PRC–(ΣOPC/MC)}/dt = >>> P , “very, very large positive -error-dot” (Very, very Less
than projected cost overtime)

The next step in the fuzzy model is to generate rule matrix for the varying conditions, so as to
come up with effective rules structure for the model.

Rule – matrix
Error
1 2 3
NC GC LC
Error - dot

4 5 6
NCR NCC LGC

7 8 9
VLGC LLG VLLG

From the rule-matrix, a rule structure can thus be generated for the fuzzy logic model as
follows:

Rule structure
IF PRC – (∑OPC/MC) = Z AND d{PRC-(∑OPC/MC)}/dt = Z THEN Output = NC
IF PRC – (∑OPC/MC) = N AND d{PRC-(∑OPC/MC)}/dt = N  THEN Output = GC
IF PRC – (∑OPC/MC) = P AND d{PRC-(∑OPC/MC)}/dt = P THEN Output = LC
IF PRC – (∑OPC/MC) = Z AND d{PRC-(∑OPC/MC)}/dt = R Z THEN Output = NCR
IF PRC – (∑OPC/MC) = Z AND d{PRC-(∑OPC/MC)}/dt = CR Z THEN Output = NCC
IF PRC – (∑OPC/MC) = N AND d{PRC-(∑OPC/MC)}/dt = >> N  THEN Output = LGC

5
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

IF PRC – (∑OPC/MC) = N AND d{PRC-(∑OPC/MC)}/dt = >>> N  THEN Output = VLGC

IF PRC – (∑OPC/MC) = P AND d{PRC-(∑OPC/MC)}/dt = >> P THEN Output = LLC
IF PRC – (∑OPC/MC) = P AND d{PRC-(∑OPC/MC)}/dt = >>> P THEN Output = VLLC

Note

Notations Meaning
∑ Summation sign
> Greater than
>> Very greater than
>>> Very very greater than
NC No change
GC Greater changer in projected cost
LC Lower/lesser change in projected cost
NCR No change remains
NCC No change continues
LGC Large greater change
VLGC Very large greater change
VLLG Very large lesser change
LLG Large lesser change

System operating rules: A System Operating rule based on the rule structure in the previous
section can thus be generated for the control of the fuzzy logic model. The operating rule is
as follows:

INPUT #1: (“Error”, Zero (Z), Negative (N), Positive (P))

INPUT #2: (“Error-dot”, Zero ( Z ), Negative ( N
 ), Positive ( P ))

CONCLUSION: (“Output”, No change (NC), Greater change (GC), Lesser change (LC),
No change remains (NCR), No change continues (NCC),
Large greater change (LGC), Very large greater change (VLGC),
Large lesser change (LLC), Very large lesser change (VLLC))

INPUT #1: System Status

Error: PRC – (∑OPC/MC)
Z = No, N = Greater, P = Less
INPUT #1: System Status
Error-dot: d(Error)/dt
Z = Not changing, N = Getting greater, P = Getting lesser
OUTPUT: Conclusion and system response
Output: N = No response, G = Greater response, L = Low response

6
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

Application: Representing the components of the fuzzy logic control model of the Gantt
charting multiple machines’ preventive maintenance activities with membership functions,
we have the following table:

Table 1: Showing relationship between fuzzy output and membership function

Level No. Interpretation Fuzzy Output Linguistic Variables
1. Optimistic Negative {(∑McOc) – (TP)c}
2. Most Likely Zero {(∑McOc) – (TP)c}
3. Pessimistic Positive {(∑McOc) – (TP)c}

where degree of relationship between fuzzy output and membership function ranges from 0 –
1.0. Illustrating the above table graphically, we have:

1.0

0.9

0.8
Membership

0.5

0.4

0.3

0.2

1 2 3 4
Note: denotes positive
denotes negative
denotes zero

The interpretation of the graph shows that:

(i) When the sum of maintenance and operating costs is higher than the total minimum
preventive maintenance cost (TP)c, the fuzzy model prompts positive.
(ii) When the sum of maintenance and operating costs are lower than (TP)c, the fuzzy
model prompts negative (optimistic output).
(iii) When the sum of maintenance and operating costs are equal to the (TP)c, the fuzzy
model prompts zero and 0 is the weakest relationship while 1.0 is the strongest
relationship.

Representing the fuzzy logic model relationship by the use of direct graph we have:
P
O
0.5
X1 X2

0.3
0.8

X3
N
7
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

Fig. 2: shows the strength of relationship within the fuzzy logic outputs: positive, zero and
negative.

Note:

X1 = Positive (High costs of maintenance and operation) (pessimistic output)

X2 = Zero (Maintenance and operating costs = minimum total production cost)
X3 = Negative (Maintenance and operating costs lower than minimum total production
cost) (optimistic output)
0.8 = Strong relationship
0.5 = Weak relationship
0.3 = Weaker relationship
This implies that it is not likely that the industry will run at high cost when sum of
maintenance and operating costs are lower than minimum total production cost i.e.
0.3
X1 X3 = Weaker relationship

P N

Also, when the sum of maintenance and operating costs is equal to the minimum total
production cost, it is easier to run at high cost compared to when the sum of maintenance and
operating cost is lower. Hence,

0.5
X1 X2 = Weak relationship

Finally, the graph implies that it is very easy to run at a costs of maintenance and operation
equal minimum total cost even though the sum of maintenance and operation costs are lower
than minimum total production costs, if care is not taken. This is illustrated from relationship
between X2 and X3 given by:

0.8
X2 X3
= Strong relationship

For more simplified graphical illustration, the following table of membership function, cost
level and cost percentage is given:

Table 2: Showing relationship between the cost level and membership function
Membership Cost Level Cost (%)
0–2 Low cost 0 – 40
2.5 – 4.5 Medium cost 41 – 50
5 – 8.5 High cost 51 – 80

8
 KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

9 – 1.0 Maximum cost 80 – 100

Fig. 3: Graphical Illustration of Table 2

1.0

0.9

0.8
Membership function

0.5

0.4

0.3

0.2
10 20 30 40 50 60 70 80 90 100
Cost (%)
To define the fuzzy logic model more clearly, a term is used to represent the procedure
followed. This term is known as “defuzzification”.

Defuzzification
Defuzzification is the term used for the procedure followed in an attempt to convert the fuzzy
set having overall conclusions to a single conclusion or value. There are various methods
used in defuzzification, the common one being the “center-of-gravity defuzzification”
method, in which the center of gravity of the overall fuzzy set conclusions is determined to
represent the desired conclusion. If a fuzzy set is represented by Z, defined over the interval
[a, b], the center-of-gravity defuzzification c is given by:
c = ∫a xz(x)dx
b

∫
b

a
z(x)dx
Based on this idea of defuzzification, an extension of fuzzy logic model is thought of in
looking into how more specific output can be achieved. This leads to further research into
neuro-fuzzy model, which intends to combine both fuzzy logic and neural networks.

A case study

TEM-D Mechanical Components Manufacturing Company (Nigeria): The above named firm
is an hypothetical organization that specializes in manufacturing of mechanical components
like bolts, nuts, screws and the likes. The major manufacturing machine is lathe machine
amongst others. A study was carried out in 2002 concerning the costs of operating and
maintaining the lathe machine in relation to the total production cost of components produced
by the machine over a period of two years. It was found that one thousand pieces each of
both bolts and nuts were produced over a period of one month at the production cost of N200
per piece. This production cost estimate per piece when calculated over a month amounted to
N200,000 for each components, giving total production cost of N400,000. It was also
discovered that some costs were incurred in maintaining the machine prior to production and
also during production. These costs are termed maintenance and operating costs respectively.
To maximize profit – TEM-D intends to sell the products at the price double the total cost of

9
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

production which implies that any extra costs (maintenance and operating costs) on total
production cost must be less than the total production cost for TEM-D to make profit. This
objective necessitates the use of a control model like fuzzy-logic control model.

Application of fuzzy-logic control model: The major parameters in the Gantt Charting
Multiple Machines’ Preventive Maintenance Activities as identified in the case study of
TEM-D manufacturing company are: Maintenance Cost, Operating Cost and Total
Production Cost. Defining these parameters in terms of fuzzy logic model, we have:

(i) Input Parameters: (a) Maintenance Cost (X)

(b) Operating Cost (Y)
(c) Total Production Cost (N400,000)
(ii) Linguistic Variables (∑XY – N400,000)
(iii) Output Parameters (a) (∑XY – N400,000) = Positive = (Pessimistic) (Pe)
Errors (b) (∑XY – N400,000) = Zero = (Most Likely) (ML)
(c) (∑XY – N400,000) = Negative = (Optimistic) (Op)

(d) d/(∑XY – N400,000)/dt = Positive = (P e )

(e) d/(∑XY – N400,000)/dt = Zero = M  ( ) Error-dot
( )
L

(f) d/(∑XY – N400,000)/dt = Negative = O 

p

Note:
Errors are output parameters within a month. Error-dot are output parameters over the period
of 2 years during which the research was completed.

A rule-matrix is thus generated from the output parameters in order to formulate a rule
structure for the fuzzy logic model given as follows:

Rule Matrix
(Error)
P Z N
(Error-dot)

1 2 3
Pe ML Op

Now formulating rule structure we have:

IF (∑XY – N400,000) = P, AND d(∑XY – N400,000)/dt = P , THEN output = Pe

IF (∑XY – N400,000) = Z, AND d(∑XY – N400,000)/dt = Z , THEN output = ML
 , THEN output = Op
IF (∑XY – N400,000) = Z, AND d(∑XY – N400,000)/dt = N

Note: The main objective of TEM-D manufacturing company is to reach out for the condition
of optimistic (Op) where (∑XY – N400,000) = Negative. A situation where the sum of the
costs of maintenance and operation are lesser than the total production cost (N400,000).

From the rule-structure, system operating rules for the sake of computation are generated as
follows:

10
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

System Operating Rules

INPUT # 1: (“Error,” Positive (P), Zero (Z), Negative (N))

INPUT # 2: (Error-dot,” Positive ( P ), Zero ( Z ), Negative ( N
 ))
CONCLUSION: (“Output”, Optimistic (Op), Pessimistic (Pe), Most Likely (ML))
INPUT #1: System Status
Error = (∑XY – N400,000)
P = Pessimistic, N = Optimistic, Z = Most Likely
Error-dot = d(∑XY – N400,000)/dt
P = Getting Pessimistic, N
 = Getting Optimistic, Z = Getting Most Likely
OUTPUT Conclusion & System Response
Output Op = Optimistic, Pe = Pessimistic, ML = Most Likely

CONCLUSIONS
As the engineering community considers approaches and models that would improve the
existing scheduling tools for both operations and maintenance, a strong case is made for the
use of fuzzy logic control model in this work. In particular, we express the mathematical
expression in the mathematical model developed by Charles-Owaba (2002) in linguistic
variables, taking in account the “error” and “error.dot” concepts in an uncertain environment.
The following observations are made in an attempt to develop and apply the fuzzy-based
model:

• The system is most desirable to operate under the condition of “Less Change” (LC) in
projected Cost (PRC) of operation and maintenance of machines, since the ultimate desire
of the industry is to minimize to the bearest minimum the costs of operation and
maintenance to achieve maximum profit.

• The fuzzy logic model gives allowance for variations in conditions, hence it is effective in
controlling the costs of operation and maintenance of machine, no matter the variations.
This accommodation of varying changes in conditions is enhanced by the fuzziness of
fuzzy logic model.

• The fuzzy logic model effectively takes care to different time periods within which
operation-maintenance activities are carried out (arrival time and post arrival time) since
it controls “error” due to changes over time.

• It, however, recommends that the operation-maintenance condition should be targeted at

achieving the least change possible PRC – (∑OPC/MC) = LC. LC – implying less
change where (∑OPC/MC) is very small. According to the fuzzy logic model, the error at
this condition is positive (P) indicating a safe condition.

In this work, four important questions readily come to the mind of a potential reader of this
article. These questions are addressed here (i) what are we going to learn from the article that
we do not know now? (ii) why is it worth knowing? (iii) how will we know that all

11
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

conclusions are valid? (iv) what does the future hold in reformulating and developing the
model?

Till date, all the articles reviewed in this work relevant to the simultaneous scheduling of
maintenance and operations have largely ignored the application of soft computing tools.
Since uncertainty exist in model application, the omission of fuzzy logic in the modeling
effort is a serious gap. This is a new dimension of treating the simultaneous scheduling of
resource-constrained maintenance and operations. It is therefore a new contribution to
knowledge in the area, and it is worthy of future extension. Since imprecision exist in real-
life measurement using the previously defined models by other authors, these uncertainties
could be tracked with a level of confidence. This is the main reason the technique is what
knowing. Based on the empirical analysis made in this paper, the results are pointers to the
validity of our conclusions. This is obviously a strong support for the conclusion made in
this work. In considering what the future holds in reformulating the model, a number of
application tools in other fields readily comes to mind. Some of the tools that will be applied
includes genetic algorithms, artificial neural network, fuzzy, neuro-genetic and a host of
statistical tools.

REFERENCES

1. Anily, S., Glass, C.A. and Hassin, R., 1998, The scheduling of maintenance service,
Discrete Applied Mathematics, 82, pp. 27-42.
2. Anily, S., Glass, C.A. and Hassin, R., 1999, Scheduling of maintenance services to three
machines, Annals of Operations Research, 86, pp. 375-391.
3. Alfares, H.K., 1999, Aircraft maintenance workforce scheduling, A case study, Journal of
Quality in Maintenance Engineering, 5(2), pp. 78-88.
4. Ashayeri, J; Teelen, A and Selen W., 1996, A production and maintenance planning model
for the process industry. International Journal of Production Research, 34 (12), pp. 3311-
3326.
5. Cheung, K.Y., Hui, C.W., 2004, Short-term site-wide maintenance scheduling,
Computers and Chemical Engineering, vol. 28, issues 1-2, pp. 91-102.
6. Charles-Owaba, O.E., 2002. "Gantt charting multiple machines’ preventive maintenance
activities", NJERD, Vol. 1, No. 1, pp. 60-67.
7. de Campos, F. C. and Belhot, R.V., 1994, Maintenance management of vehicle fleets: a
review. Gestao & Producao (Brazil), 1(2), p. 171-189.
8. Dieulle, L., Berenguer, C., Grall, A., Roussignol, M., 2003, Sequential condition-based
maintenance scheduling for a deteriorating system, European Journal of Operational
Research, vol. 150, issue 2, pp. 451-46
9. Duffuaa, S.O. and Ben-Daya, M., 1994, An extended model for the joint overhaul
scheduling problems, International Journal of Operations Management, 14(7), pp. 37-43.
10. El-Sharkh, M.Y., El Keib, A.A., 2000, An evolutionary programming-based Solution
methodology for power generation and transmission maintenance scheduling,
Transportation Research part A: Policy and Practice, vol. 36, issue 5, pp. 453-482.
11. El-Sharkh, M.Y., El-Keib, A.A., 2003, A fuzzy evolutionary programming-based solution
methodology for security-constrained generation maintenance scheduling, Electric Power
Systems Research, vol. 67, issues 1, pp. 67-72.
12. Gunn, E.A. and Lee, B.P., 1991, Lagrangian relaxation and integer programming
heuristics for vehicle fleet maintenance scheduling, Paper Presented at International
Symposium on Mathematical Programming, Amsterdam.

12
KATHMANDU UNIVERSITY JOURNAL OF SCIENCE, ENGINEERING AND TECHNOLOGY
VOL.I, No.III, JANUARY, 2007.

13. Kobbacy, K. A. H., Fawzi, B. B., Percy, D. F., Ascher, H. E., 1997, A full history
proportional hazards model for preventive maintenance scheduling, Quality and
Reliability Engineering, 13, pp. 187-198.
14. Lake, M. and Ferreira, L., 2001, Considering the risk of delays in scheduling railway
track maintenance, Proceedings of the 7th International Heavy Haul Conference,
International Heavy Haul Association Inc., Virginia Beach, VA, USA, pp. 367-372.
15. Lake, M., Ferreira, L. and Murray, M., 2000, Minimising costs in scheduling railway
track maintenance, Chapter 13, pp. 895-902. In: Computer in railways
VII, Allan, Hill, Brebbia, Scuitto and Sone (eds), WIT Press, Southampton.
16. Lee, P., 1991, Integer programming approaches to vehicle fleet maintenance scheduling,
Unpublished M.A.Sc Degree Thesis, Department of Industrial Engineering, Dalhousie
University.
17. Oke S.A., 2004a, A re-examination of the optimal gantt charting model for maintenance
scheduling, Proceedings of MIMAR 2004, the 5th IMA International Conference, 5-7
April 2004, Salford, United Kingdom, pp. 219-224, ISBN number 0905091159.
18. Oke S. A., 2004b, Maintenance scheduling: Description, status, and future
directions, South African Journal of Industrial Engineering, 15(1), pp. 101-117.
19. Olorunniwo F.O; Izuchukwu A., 1991, Scheduling imperfect preventive and overhaul
maintenance, International Journal of Quality & Reliability Management, 8(4).
20. Ram, B. and Olumolade, M., 1987, Preventive maintenance scheduling in the presence of
a production plan, Production and Inventory Management, 8, pp. 81-89.
21. Stremel, J.P., 1981, Maintenance scheduling for generation system planning, IEEE
Transactions on Power Apparatus and Systems, PAS-100(3), pp. 4010-4019.
22. Stremel, J.P., Jenkins, R.T., 1981, Maintenance scheduling under uncertainty, IEEE
Transactions on Power Apparatus and Systems, PAS-100(2), pp. 460-465.
23. Wang Y. A., 2002, Hybrid genetic algorithm-simulation methodology for pavement
maintenance scheduling, Unpublished M. Eng. Dissertation, Department of Industrial
Engineering, National University of Singapore, Singapore.
24. Walker, L., Bryan, U. and Turner, K., 2001a, Scheduling preventive maintenance of
transmission circuits, 4th IMA Conference on Modelling in Industrial Maintenance and
Reliability Decision Support in the New Millennium, April, 2001, UK.
25. Zurn, H.H., Quintana, V.H., 1977, Several objective criteria for optimal generator
preventive maintenance scheduling, IEEE Transactions on Power Apparatus and Systems,
PAS-96(3), pp. 984-922.

26. Zurn, H.H. and Quintana, V.H. 1975, Generator maintenance scheduling via successive
approximations dynamic programming. IEEE Transactions on Power Apparatus and
Systems, 94(2), pp. 665-671.

13