Failure-Based Maintenance Decision Support System Using Analytical Hierarchical Process

8/6/2019 Failure-Based Maintenance Decision Support System Using Analytical Hierarchical Process

http://slidepdf.com/reader/full/failure-based-maintenance-decision-support-system-using-analytical-hierarchical 1/9

International J ournal of Advanced C omputer Science, Vol. 1, No. 1, Pp. 1-9, Jul. 2011 .

Manuscript Received:26, June, 2011

Revised:21, July, 2011

Accepted:30, July, 2011

Published: 10, August, 2011

Keywords Analytical hierarchical

process, pair-wisecomparison,maintenancecontractor benchmarking,maintenancemanagement

system,decision

support model

Abstract This research based on thefact that the investment of informationsystem is needed for supporting dailymaintenance management in industries.Currently, maintenance department islacking of appropriate Decision SupportSystem (DSS) to monitor contractors andtechnicians in the field work. Althoughmany models with varieties optimizationtechniques have been proposed, there arestill limited researches to embed them with

computerized maintenance managementsystem. The objective of this paper is topropose a structured DSS aid maintenancemanagement to benchmark contractors inthe field work. The proposed DSS based onthe given goals, criteria, sub-criteria,alternatives and constraints usingAnalytical Hierarchical Process method(AHP). The method is suitable formaintenance DSS with multiplemaintenance criteria’s and theiralternatives. The system is developed usingevidence from an empirical study. Thepaper aims to demonstrate how the model

can help in solving such decisions inpractice. We collected real data from one of the food processing factories in Malaysiaand test DSS for decision making. DSS isdeveloped with AHP to benchmark fourmain contractors in failure-basedmaintenance (FBM). Based on the resultusing our system, contractors are ranked:C, B, A, D in ascending order for overallperformance in 2006. FBM managementwith decision analysis using AHP is ourcontribution in this paper for advancedcomputerized system.

1. IntroductionBasically industrial organizations have been divided in

operative functions such as marketing, planning, production,maintenance, purchasing, procurement, etc.

All of these functional department use machines thatrequire proper maintenance activities. The lack of capitaland less effective in carrying out the maintenance of

This research work supported by the Ministry of Science and Technology,Malaysia. M. A. Burhanuddin is a senior lecturer in Universiti TeknikalMalaysia Melaka, Sami M. Halawani and A. R. Ahmad are associate

professors in King Abdulaziz University Rabigh, Saudi Arabia, ZulkifliTahir is a lecturer in Universitas Hasanuddin, Indonesia.([email protected], {halawani,abinahmad}@kau.edu.sa,[email protected])

production process are the main problems that appeared inindustries. In Malaysia, although the government given a lotof capital assistance, due to lack of information abouteffective maintenance, has made the production inindustries still cannot reach the expectations [1]. Past work shown that maintaining own machines without outsourcingwill save cost in a long term. However, more industries tendto outsource their maintenance function at present, wherethey just follow the maintenance advice from thecontractors to perform maintenance activities. This is

because it is easier to manage, but this will increasemaintenance cost and utilizing the contractors always anissue. To overcome this problem, Decision Support System(DSS) can be used by mining historical data, controlmaintenance activities and predicts future maintenancestrategies.

Recently, many models with varieties optimization andtechniques have been proposed ([2], [3] and [4]). However,there are limited actions to be linked into the actualindustrial maintenance process [5].

In order to increase the effectiveness of the units,computerized system like DSS needed to simplify theanalyzing process and to reduce the time needed for maintenance decision. The purpose of this study is todemonstrate DSS with Analytic Hierarchy Process (AHP)analysis. In AHP, the goal or the decision that is suggestedfrom any maintenance problem can be determined fromselected alternatives.

2. Literature on MCDM ModelsAt present, there is a boom in maintenance management

research, where a large amount of publications haveappeared on maintenance performance measures andtechnicians key performance index. Literature search andstudies about the models and proposed techniques are

summarized in the following paragraphs.In early 1986, Timmerman [6] proposed linear

weighting models in which vendors are rated on severalcriteria and in which these ratings are combined into asingle score. These models include the categorical, theweighted point [6] and the analytical hierarchical process[7]. The major limitation of this approach is that it isdifficult to effectively take qualitative evaluation criteriainto consideration. Total cost approaches attempt toquantify all costs related to the selection of a contractor inmonetary units. This approach includes cost ratio [6] andtotal cost of ownership [8].

Petroni and Braglia [9] discuss the principal componentanalysis method which is a multiobjective approach tocontractor selection that attempts to provide a useful

Failure-based Maintenance Decision Support Systemusing Analytical Hierarchical Process

M. A. Burhanuddin, Sami M. Halawani, A.R. Ahmad, & Zulkifli Tahir




International J ournal P ublishers G roup ( IJPG ) ©

2

decision support system for a purchasing manager facedwith multiple contractors and trade-offs such as price,delivery, reliability, and product quality. The major limitation of this approach is it requires the knowledge of advanced statistical techniques. Wei et al. [10] in their paper discuss about the neural network for the supplier selection.When [10] compared conventional models for DSS withneural networks, they discovered that neural networks savea lot of time and money during system development. Thesupplier-selecting system includes two functions: one is thefunction measuring and evaluating performance of

purchasing (quality, quantity, timing, price and costs) andstoring the evaluation in a database to provide data sourcesto neural network. The other is the function using the neuralnetwork to select suppliers. This method incorporatesqualitative and quantitative criteria. However, the weaknessof this method is that it demands software and requires a

qualified personnel expert on this artificial neural networksfield.

One of the most common problems in manyengineering and business applications is how to evaluate aset of alternatives in terms of a set of decision criteria. For example, let someone intends to set up computer servers.There are a number of different configurations available tochoose from. The different operating systems are thealternatives. Here, a decision should also consider cost and

performance characteristics, such as processing unit speed,memory capacity, network bandwidth, number of clients,etc. Alternatives of software, maintenance, expendabilityshould be considered too. These may be some of thedecision criteria for this case, where the criteria may vary

based on different purpose of the servers. AHP is one of themultiple criteria decision making models, always used for

problem-solving to determine the best alternative andsolution.

AHP developed by [11] as mathematical decisionmaking model to solve complex linear algebra problemswhen there are multiple objectives or criteria to beconsidered. It’s requires the decision makers to provide

judgments about the relative importance criterion for eachdecision alternatives [12]. Reference [13] has used AHP to

justify the Total Productive Maintenance, while [14] havedescribed the application of AHP to selecting the bestmaintenance strategies for an important Italian oil refinery.Moreover, reference [1] has been evaluated the informationservice quality of ten primary high tech industryinformation center web portals in China using AHP. Then,AHP also has been used to evaluate the call center servicequality in Taiwan telecommunication industries by [2].

There are several steps to implement the AHP model [3].The first step is describing the maintenance problem into anAHP decision hierarchy. The hierarchy can be visualized asa diagram in Fig. 1. It consists of an overall goal at the top,a group of options or alternatives for reaching the goal atthe bottom, and a group of factors or criteria filling up at themiddle; relate the alternatives to the goal. In most cases, thecriteria and then divided into sub-criteria in some degree

based on the needs of the problem. Clearly, the AHP is most

efficient applied when the total number of criterions andalternatives is not excessive [11].

Fig. 1. The AHP Hierarchy

The AHP concept establishes the priorities in eachelement in hierarchy by make a pair wise comparison of thecriterions and alternatives in every level. The comparisonsare predefined on ratio scale as listed in Table 1.

Table 1. Scale of the ImportanceIntensity of Importance Value Description Explanation

1 Criteria i and j, areequal importance

Two activities contributeequally to the objective

3Criteria i is weaklymore importantthan j

Experience and judgmentslightly favor oneactivity over another

5

Criteria i is strongly

more importantthan j

Experience and judgment

strongly favor oneactivity over another

7Criteria i is verystrongly moreimportant than j

An activity is stronglyfavored and itsdominance demonstratedin practice

9Criteria i isabsolutely moreimportant than j

The evidence favoringone activity over another is the highest possibleorder of affirmation

2, 4, 6, 8

Intermediate values between the twoadjacent values. If criterion i has oneof the abovenon-zero numbersassigned to it whencompared withcriterion j. then jhas the reciprocalvalue whencompared with i

When a compromise in judgment is needed

e. g. if i = 4, j = ¼

The comparisons are made using the judgments of thefactors based on data obtained in the SMIs or from theknowledge and experience of the maintenance persons, i.e.technicians, managers, or other experts in the maintenancedepartment. There are many situations where the judgments

are close or tied in measurement and the comparison must be made between one-to-nine ratio scales. For examplethere are comparisons to be made between 1 and 2, such as



Burhanuddin et al .: Failure-based Maintenance Decision Support System using Analytical Hierarchical Process


3

1.1, 1.2, 1.3, …, 1.9.The example of comparison formula is shown in the

Equ. 1 below. It is a square matrix with as many rows (andcolumns) as there are criteria connected to the goal. Thenumbers in this matrix express the intensity of dominance

of the criterion in the column heading over the criteria in therow heading. In many research, the ratio scale have beenused, the matrix is reciprocal which mean that the numbers,which are symmetric which respect to the diagonal, areinverses of one another, a ij =

jia

1 . If one criterion is judged

to be three times more important than the others, then theothers is as important when compared like the first.

(Equ. 1)In general,

2)1(nn comparisons are needed if n is the

number of element being compared in the triangle above thediagonal of ones. These comparisons show in bolt variablesin Equ. 1.

Approximating the weight vector in the matrix A, with irow and j column, takes from illustrated below, where wi >0 for i = 1, …, n denotes the weight of objective i. The nextstep is the calculation of a list of the relative weights of thecriteria under consideration. This requires normalization of each column j in A, such that O j aij = 1.

(Equ. 2)

For each row i in the resulting matrix above, theaverage value is computed by:

(Equ. 3)Where w i is the weight of criterion i in the weight

vector, w = (w 1, w 2, …, w n) recovered from matrix A, withn criteria, by finding a non-trivial solution to a set of nequation with n unknowns.

Finally, given a decision matrix, the final priorities,denoted by A i

AHP , of the alternatives in term of all thecriteria combined are determined as:

(Equ. 4)Reference [15] discussed the contractor selection for the

telecommunication systems and based on the proposedmodel, they proved that the time taken to select the

contractor has been reduced.Reference [16] provided a solution to perform a

multivariate design and multiple criteria analysis of alternate alternatives based on the enormous amount of information. They develop multivariate design method and

multiple criteria for a building refurbishment’s analysis in public building of Vilnius Gediminas Technical University.They use matrix Weightage analysis for decision making.Reference [17] identified optimal maintenance planningarrangement of maintenance schedules for PM byconsidering few criteria: availability, maintenance cost, lifecycle costs and tolerance level. They appliedMultiple-objective Genetic algorithm MATLAB toolbox in

paper production, nuclear production plant. [18] usedmultiple criteria evaluation of multi-family apartment

block's to benchmark maintenance contractors. They builtthe model for maintenance contractor evaluation and thedetermination of its selection. Later, [19] applied [18]’smodel in Lithuanian case study. Reference [20] applied [21]and [22] approaches by implementing Decision MakingGrid (DMG) in a food processing factory. They discoveredthat the DMG analysis creates more opportunities in CMMSto be transparent to the entire production workforces,instead of only the maintenance team. The increased usageof data mining and DMG analysis in CMMS by personneloutside maintenance function may have a good potential toimprove maintenance strategies and equipment utilization.Reference [20] recommended that upgrading operators’technical skills in repairing the machines in the FTM1,FTM2 and SLU regions will reduce the cost of maintenancein the long-term, because operators are able to fix the

problem without any escalation to the maintenance team.They recommended AHP model to benchmark techniciansas future research direction.

3. The Proposed DSS DesignThere are many researchers who have studied decision

support techniques and apply it in the machineriesmaintenance management area. Maintenance strategies can

be explained as a collection of maintenance operation thatgives best effort to benefit with least cost [23]. Eachmaintenance operation have specific scenario to beimplemented to get the best result with minimum cost [24].

In this research, the AHP has used to help decision makersto make proper evaluations and relatively accurate decisions.The example case is given in Fig. 2.

Fig. 2 shows the hierarchy framework of contractor selection based on specific criterions and alternatives. Inthis research, the data was collected from a food processingfactory in Malaysia.The hierarchy is used for conducting preliminary analysis indomain of contractor selection. Basically, the contractor requirement can be decomposed into several criterions andsub-criterions in different level of uncertainty and ambiguity.Analysis can be performed at each level independently, butlinked and cumulated at the higher levels in the hierarchy.

Decisions and judgments are made in each level of structure,and finally aggregated to produce decision in the top of hierarchy.





4

Fig. 2. Contractors Selection

In conducting pair wise comparisons between contractor selections, the decision makers sequentially compares two

criterions and alternatives. In the criterions level in Fig. 2,the system needs 6 pair wise comparisons input. In subcriterions, the system need 1 pair wise comparison for technical performance sub criterion and need 3 pair wisecomparison for business principle sub-criterion. Moreover,

6 pair wise comparison are needed in alternatives levellinked for every higher level in hierarchy. Base on our interview, all of the pair wise comparison calculations have

been inputted and generated [4].In our maintenance system, the database relational

design using AHP decision model is shown in Fig. 3.The AHP database design is shown in Fig. 4.For all pair wise comparisons, we construct pair wise

comparison matrices using Equ. 1. After complete,normalization matrix are produced by divided the number of

Fig. 3. Database Relation

Fig. 4. AHP Database Design





5

matrix by their respective column using Equ. 2. Then todetermine the priorities, we simply find the average of thevarious rows from normalization matrix using Equ. 3. Theresult shows in Table 2 for every criterions andsub-criterions, those are the priority evaluation weights of

Technical Performance (TP), Business Principles (BP),Experience in Maintenance (EM), Action and Presence inTroubleshooting Activities (AP), Quality of MaintenanceWork (QW), Tool and Equipment (TE), Enthusiasm inProblem Solving (EP), Repair Time (RT) and ResponseTime (RnT).

The activity diagram on AHP goal is given in Fig. 5.

Fig. 5. Activity Diagram of AHP Goal

Finally, to get the overall ranking, the priority evaluationweights are multiplied in each table using Equ. 4. The dataevaluations for four different contractors are summarized oncolumn TOTAL in Table 2. The data shows the weight of contractor C received the highest final ranking as the bestcontractor in the factory.

As discussed, solving AHP problems can involve alarge number of calculations. Therefore, the AHP

program as Decision Support System (DSS) is applicablefor critical decision support and elaborated by [5].

4. DSS ImplementationIn this section, AHP DSS application development

and implementation have been outlined. DSS at leastmust have required data in response to the maintenanceevent including breakdown machine, downtime historyand all the parameters such as contractor selection.Furthermore, the system must have the capability to

show the appropriate maintenance strategies andmaintenance decision goal base on the data inputted.Hence, our system includes the following modules.

(i) Maintenance data to get the appropriate datainput; and

(ii) Analytic Hierarchy Process calculation formulafor analysis on decision goal for eachmaintenance alternatives and criterions.

The program must be running some form of webserver software such as Apache and Windows Server, andmust have a continuous connection to the internet asshown in Fig. 6. The server must also have HypertextPreprocessor (PHP) installed on it so that program will

be able to work. All of the program files must be copiedover into the web domain to function properly. For database, server must have MySQL database installed.All of the data can be uploaded in that database server.

Once the installation is complete the site will be fullyaccessible from World Wide Web at the domain addressthat it is being hosted from anywhere.

Fig. 6. Web DSS Application Concept

Use case diagram for AHP development is given in Fig.7.

Best Contractor Selection MaintenanceTOTAL

ContractorsTP BP

RT RnTEM AP QW TE EP

A 0.074 0.030 0.007 0.009 0.011 0.047 0.068 0.245B 0.049 0.044 0.010 0.013 0.027 0.083 0.058 0.284C 0.099 0.059 0.014 0.004 0.005 0.081 0.041 0.304D 0.025 0.015 0.003 0.017 0.011 0.049 0.047 0.167

TOTAL

0.247 0.148 0.034 0.043 0.0530.260 0.214

1.0000.396 0.1301.000

Table 2. Estimation of Contractors with AHP





6

Fig. 7. Use Case Diagram AHP

We use several tools to build the DSS applications. For server and clients hardware, we use Hewlett-Packard (HP)sw4400 Work Station Intel(R) Pentium(R) D CPU 3.40GHz 1GB of RAM and Monitor HP L1506 and MIMOS

Intel(R) Pentium(R) 4 CPU 2.80 GHz, 248 MB of RAMand Monitor MIMOS. Microsoft Windows XP has beenused as Operating System. To develop PHP code,Macromedia Dreamweaver has been used. Apachedistribution containing MySQL and PHP using XAMPPWIN 32 are used as web server includes User Interface (UI),PHP interpreter and database.

5. Graphical User InterfaceWe wrote AHP formulae in our DSS to estimate goal

from multiple alternatives and multiple criterion. Thedecision maker should be authorized personnel to executethe next process in sequent from AHP goal until AHPdecision reached. In our program, the best goal only can becomputed after all alternatives captured. Main interface of FBM management system is shown in Fig. 8.

Fig. 8. Main Interface of FBM System

One of our criterion input interface is shown in Fig. 9.The program designed into a display screen with a user can

interact using a computer input devices either keyboard or mouse.

Fig. 9. Maintenance DSS Criteria Input

One of our sub-criterion input interface is shown in Fig.10.

Fig. 10. AHP Sub-criterion Input

AHP Alternative Comparison input interface is given in Fig.11.

Fig. 11. AHP Alternative Comparison Interface





7

Our experimental result using AHP method is given in Table3.

We justify our system with the Expert Choice Softwareto compare and validate our result. It shows similar resultsin Expert Choice software, compared to our DSS. Our comparison study using AHP method in our system is givenin Table 4.

Our system is better compared to the Expert Choicesoftware as all data is captured in the database and users areable to recall the information anytime. Moreover, our web

based DSS is the best as it simplified, paperless and reducethe data acquisition time compared to other stand alonesystem or manual record-keeping system [5]. This systemalso provided the maintenance plan with the application for

analysis and decision support that often ignored by many proposed computerized base maintenance managements for industries.

A sensitivity graph is given in Fig. 12 from AHP analysis.The x-axis shows the available objectives or main criteriaand the y-axis shows the satisfaction of the contractor’s

performance in percentage towards each criteria. The performance sensitivity graph of contractors shows therelative importance of each of the criteria as bars. Therelative preference for each contractor with respect to eachcriterion is given as the intersection of the contractors’, witha given vertical line for each criterion.

Fig. 12. Contractor Sensitivity Graph

Note that the points are connected using lines, just tovisualize the intersection among contractors, but actuallythere are no connections to each other. Overall, contractorsare ranked in the ascending sorted list: C, B, A and Drespectively.

6. ConclusionThis paper demonstrates DSS to analyze the maintenancedata in FBM. AHP is the maintenance optimizationtechnique using linear algebra and Eigen values calculation,

been used in our system. This AHP method is able toestimate the risk factors and provide maintenance crucialdecisions for industries. In the proposed system, it is

possible for users to key in all relevant data in computerizedmaintenance management system. Next, they can update itsrelative importance variables and weights to evaluate thealternatives. We have shown how AHP is used to analyzeand rank contractors. Our DSS is able to provide a clear rationale to get the best decisions. In our case studies,analyses and the result for a precise decision-making

process is given. The result can tell decision-makers in twoways: firstly, to know the alternatives and their flexibility of choices; and, secondly, for forecasting, by evaluating and

benchmarking contractors for maintenance breakthrough performance. This paper provides salient contributions tothe domain of knowledge in advanced computerized system,to develop DSS to benchmark maintenance contractors for maintenance job using the mathematical models.

Next, the analysis is also carried out with theimplementation of the DSS programs and has inspired us to

update the AHP model program with the addition of intelligent logic and the suitability of algorithms concept.Future research to hybrid AHP with other maintenancereliability techniques in computerized maintenancemanagement system could aid maintenance managers toderive important maintenance strategies.

AcknowledgmentThe authors would like to thank Faculty of Computing

and Information Technology Rabigh, King AbdulazizUniversity Rabigh for financial support and facilities. Alsothanks to Faculty of Information and CommunicationTechnology, Universiti Teknikal Malaysia Melaka for grantthe main author for the attachment program in KingAbdulaziz University.

References[1] M. Guo, & Y. Zhao, “AHP based evaluation of information

service quality of information center web portals in high-techindustries,” (2009) IEEE on management and service science ,vol. 9, pp. 1-4.

[2] W.H. Chou, & C.H. Chen, “Using AHP to evaluate the callcentre service quality in telecomunication industry,” (2009)

IEEE on web information systems and mining , pp. 823-827.[3] T.L. Saaty, “Decision making with the analytic hierarchy

process,” (2008) International Journal of Services Sciences, Inderscience , vol. 1, no. 1, pp. 83-98.[4] M.A. Burhanuddin, “Decision Support Model in

Comparison of AHP Decision Analysis AHP Alternative FBM System Expert Choice ErrorContractor A 0.245 0.246 0.001Contractor B 0.284 0.283 0.001Contractor C 0.304 0.303 0.001Contractor D 0.167 0.167 0

Table 4. AHP Comparison Analysis

FBM AHP Decision

AHP Alternative WeightContractor C 0.304Contractor B 0.284Contractor A 0.245Contractor D 0.167

Table 3. AHP Decision Result





8

Failure-Based Computerized Maintenance ManagementSystem for Small and Medium Industries,” (2009) Thesis of

Doctor of Philosophy in Computer Science , UniversitiTeknologi Malaysia.

[5] Z. Tahir, “A Hybrid Maintenance Management Model InDecision Support System For Small And Medium FoodProcessing Industries,” (2010) Thesis of Master of Science in

Information and Communication Technology , UniversitiTeknikal Malaysia Melaka.

[6] E. Timmerman, “An approach to vendor performanceevaluation,” (1986) Journal of Supply Chain Management ,vol. 22, no. 4, pp. 2-8.

[7] R.L. Nydick & R.P. Hill, “Using the analytic hierarchy pprocess to structure the supplier selection procedure,” (1992) International Journal Purchase Matter Management , vol. 28,no. 2, pp. 31-36.

[8] L.M. Ellram, “Total cost of ownership: an analysis approachfor purchasing,” (1995) International Journal Phys Distrib

Logistic , vol. 25, no. 8, pp. 163-184.[9] A. Petroni & M. Braglia, “Vendor selection using principal

component analysis,” (2000) Supply Chain Management Journal , vol. 36, no. 2, pp. 63-69.

[10] S.Y. Wei, Z. Jinlong, & L.I. Zhicheng, “A supplier-selectingsystem using a neural network,” (1997) IEEE International Conference on Intelligent Processing Systems , vol. 1, pp.468-471.

[11] T.L. Saaty, “The Analytic Hierarchy Process,” (1980)McGraw-Hill , New York.

[12] R. Kodali, “Qualification of TPM benefits through AHPmodel,” (2001) Productivity , vol. 42, no. 2, pp 265-73.

[13] R. Kodali, & S. Chandra, “Analytic hierarchy process of justification of total productive maintenance,” (2001) Production Planning and Control , vol. 12, no. 7, pp. 695-705.

[14] M. Bevilacqua, & M. Braglia, “The analytic hierarchy process

applied to maintenance strategy selection,” (2000) Reliability Engineering and System Safety , vol. 70, no. 1, pp. 71-83.[15] M.C.Y. Tam & V.M.R. Tummala, “An application of the AHP

in vendor selection of a telecommunications system,” (2001)Omega International Journal of Management Science , vol. 29,no. 2, pp. 171-182.

[16] S. Ana, C. Sofia, M. Sebastian, & F.V. Jose, “Addressingimperfect maintenance modelling uncertainty inunavailability and cost based optimization,” (2007)

Reliability Engineering and System Safety , vol. 10, pp. 1-11.[17] S. Anish, K. Dinesh, & K. Pradeep, “Planning and optimizing

the maintenance of paper production systems in a paper plant,”(2008) Computers and Industrial Engineering , vol. 10, pp.1-13.

[18] E.K. Zavadskas, & T. Vilutien W, “A multiple criteriaevaluation of multi-family apartment block's maintenancecontractors,” (2006) Building and Environment , vol. 41, no. 5,

pp. 621-632.[19] E.K. Zavadskas, & T. Vilutien W, “Multicriteria Evaluation of

Apartment Blocks Maintenance Contractors: Lithuanian CaseStudy,” (2009) International Journal of Strategic PropertyManagement , vol. 13, no. 4, pp. 319-338.

[20] M.A Burhanuddin, A.R. Ahmad, & M.I. Desa, “MaintenanceDecision Making Using Multiple Criteria Analysis for Smalland Medium Industries,” (2007) Journal of InformationTechnology , University Technology Malaysia, vol. 19, no. 2,

pp. 35-46.[21] O. Fernandez, R. Walmsley, D.J. Petty, & A.W. Labib, “A

decision support maintenance management system

development and implementation,” (2003) International Journal of Quality and Reliability Management , vol. 20, no.8, pp. 965-979.

[22] A.W. Labib, “World-class maintenance using a computerizedmaintenance management system,” (1998) Journal of Qualityin Maintenance Engineering , vol. 4, no. 1, pp. 66-75.

[23] J.F. Villanueva, A.I. Sanchez, S. Carlos, & S. Martorell,“Genetic algorithm-based optimization of testing andmaintenance under uncertain unavailability and costestimation: A survey of strategies for harmonizing evolutionand accuracy,” (2008) Reliability Engineering and SystemSafety, vol. 10, pp. 1-12.

[24] R. Pascual, V. Meruane, & P.A. Rey, “On the effect of downtime costs and budget constraint on preventive andreplacement policies,” (2008) Reliability Engineering and System Safety , vol. 93, pp. 144-151.

Dr. Burhanuddin MohdAboobaider obtained the M.Sc.degree in Mathematics in 2004 fromthe Universiti Sains Malaysia. He

received the B.Sc. degree inComputer and Ph.D. degree inComputer Science from theUniversiti Teknologi Malaysia in

2002 and 2009 respectively. He is the author or co-author of more than thirty national and international papers andalso collaborated in several research projects. He has beenworking with Universiti Teknikal Malaysia Melaka since2004 with Computer Vision and Robotic Research Center Group. He is an Assistant Professor and currently, onattachment at the Faculty of Computing and InformationTechnology in Rabigh, King Abdulaziz University Rabigh,Kingdom of Saudi Arabia since 2010. His current research

interests include multiple criteria decision making modelsand decision support system.

Dr. Sami M. Halawani received theM.S. degree in Computer Sciencefrom the University of Miami, USAin 1987. He received theProfessional Applied EngineeringCertificate from The GeorgeWashington University, USA in1992. He earned the Ph.D. degree in

Information Technology from the George MasonUniversity, USA in 1996. He is an Associate Professor and

actively collaborated in several research projects. Currently,he is the Dean of the Faculty of Computing and InformationTechnology Rabigh, King Abdulaziz University, Kingdomof Saudi Arabia. He has published many papers journals,conference, proceedings as author and co-author.

Dr. Ab Rahman Ahmad receivedthe M.Sc degree in Optimization andComputing, and Ph.D in NumericalAnalysis from the LoughboroughUniversity of Technology, England,in 1985 and 1993 respectively. Hehas been working at Universiti

Teknologi Malaysia since 1978 andhas been promoted to Associate Professor in 1996. He is aChairman of Information Systems Department and





9

currently on attachment at Faculty of Computing andInformation Technology in Rabigh, King AbdulazizUniversity Rabigh, Kingdom of Saudi Arabia since 2009.He has published many papers journals, conference,

proceedings as author and co-author.

Zulkifli Tahir was born inMakassar, South Sulawesi,Indonesia in 1984. He received theB.Sc. degree in Informatics fromInstitut Teknologi Telkom,Bandung, Indonesia in 2006 andM.Sc. degree in Information andCommunication Technology from

Universiti Teknikal Malaysia Melaka, Malaysia in 2010.Currently, he is a lecturer in Universitas Hasanuddin,Makassar, Indonesia. His research interests includedecision support system, industrial computing, artificial

intelligent, network and multimedia.

Documents

Failure-Based Maintenance Decision Support System Using Analytical Hierarchical Process