Welcome to Scribd, the world's digital library. Read, publish, and share books and documents. See more
Download
Standard view
Full view
of .
Save to My Library
Look up keyword
Like this
2Activity
0 of .
Results for:
No results containing your search query
P. 1
Fuzzy-Based Dynamic Rough Set Resource Discovery According to User Preferences in Grid Environment

Fuzzy-Based Dynamic Rough Set Resource Discovery According to User Preferences in Grid Environment

Ratings: (0)|Views: 585|Likes:
Published by ijcsis
Grid environment is a service oriented infrastructure in which many heterogeneous resources participate to provide high performance computation. One of the bug issues in the grid environment is the uncertainty among registered resources. Furthermore, in an environment such as grid dynamicity is considered as a crucial issue which must be dealt with. Using dynamic rough set theory to deal with uncertainty and dynamicity has shown good results. Addition to this, compounding this theory with Fuzzy system will improve its efficiency by applying fuzzy system in matchmaking phase. In this work we propose a solution, called Fuzzy-Based Dynamic Rough Set Resource Discovery (FDRSRD), in which dynamic Rough set theory is used in order to deal with uncertainty and fuzzy system for resource matchmaking according to the user preferences. In matchmaking phase, in order to improve accuracy and speedup, fuzzy system is used to rank resources. We also report the result of the solution obtained from the simulation in GridSim simulator along with FuzzyJ Toolkit. The comparison has been made between FDRSRD, Dynamic and Classical Rough Set based algorithms. FDRSRD shows much better precision and more speed for the cases with uncertainty in a dynamic system such as the grid rather than the two other algorithms.
Grid environment is a service oriented infrastructure in which many heterogeneous resources participate to provide high performance computation. One of the bug issues in the grid environment is the uncertainty among registered resources. Furthermore, in an environment such as grid dynamicity is considered as a crucial issue which must be dealt with. Using dynamic rough set theory to deal with uncertainty and dynamicity has shown good results. Addition to this, compounding this theory with Fuzzy system will improve its efficiency by applying fuzzy system in matchmaking phase. In this work we propose a solution, called Fuzzy-Based Dynamic Rough Set Resource Discovery (FDRSRD), in which dynamic Rough set theory is used in order to deal with uncertainty and fuzzy system for resource matchmaking according to the user preferences. In matchmaking phase, in order to improve accuracy and speedup, fuzzy system is used to rank resources. We also report the result of the solution obtained from the simulation in GridSim simulator along with FuzzyJ Toolkit. The comparison has been made between FDRSRD, Dynamic and Classical Rough Set based algorithms. FDRSRD shows much better precision and more speed for the cases with uncertainty in a dynamic system such as the grid rather than the two other algorithms.

More info:

Published by: ijcsis on Sep 05, 2010
Copyright:Attribution Non-commercial

Availability:

Read on Scribd mobile: iPhone, iPad and Android.
download as PDF, TXT or read online from Scribd
See more
See less

06/12/2011

pdf

text

original

 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 5, August 2010
Fuzzy-Based Dynamic Rough Set ResourceDiscovery According to User Preferences in GridEnvironment
Iraj Ataollahi
Computer Engineering DepartmentIran University of Science and TechnologyTehran, Iranir_ataollahi@mail.iust.ac.ir
Mahdi Bakhshi
Department of ComputerIslamic Azad University, Shahrbabak BranchShahrbabak, Iranmah.bakhshi@yahoo.com
Morteza Analoui
Computer Engineering DepartmentIran University of Science and TechnologyTehran, Irananaloui@iust.ac.ir 
 Abstract
—Grid environment is a service oriented infrastructurein which many heterogeneous resources participate to providehigh performance computation. One of the bug issues in the gridenvironment is the uncertainty among registered resources.Furthermore, in an environment such as grid dynamicity isconsidered as a crucial issue which must be dealt with. Usingdynamic rough set theory to deal with uncertainty anddynamicity has shown good results. Addition to this,compounding this theory with Fuzzy system will improve itsefficiency by applying fuzzy system in matchmaking phase. Inthis work we propose a solution, called Fuzzy-Based DynamicRough Set Resource Discovery (FDRSRD), in which dynamicRough set theory is used in order to deal with uncertainty andfuzzy system for resource matchmaking according to the userpreferences. In matchmaking phase, in order to improveaccuracy and speedup, fuzzy system is used to rank resources.We also report the result of the solution obtained from thesimulation in GridSim simulator along with FuzzyJ Toolkit. Thecomparison has been made between FDRSRD, Dynamic andClassical Rough Set based algorithms. FDRSRD shows muchbetter precision and more speed for the cases with uncertainty ina dynamic system such as the grid rather than the two otheralgorithms.
 Keywords-
 
Grid; Rough Set; Dynamic Rough Set; Resource Discovery; Ontology; Fuzzy System; User Preferences
I.
 
I
NTRODUCTION
 Nowadays, Grid is considered as a service-orientedcomputing infrastructure [1]. Open Grid Services Architecture(OGSA) [2] has been used for dealing with service-orientedproblem [3]. OGSA has been improved by Global Grid Forum.Many resources such as workstations, clusters, andmainframes with various properties such as main memory,CPU speed, bandwidth, virtual memory, hard disk, cost, andresponse time etc are joining and leaving the grid environment.On the Other hand many users want to use these resources torun their jobs with different requirements. But there are alwaysdifferences between which a user requested and which havebeen registered in the Grid Information Server (GIS). This maylead to uncertainty in GIS data base which will be resulted in tothe reduction of the precision of the matchmaking algorithms.To solve this vagueness and uncertainty we use rough settheory, proposed by Z. Pawlak in 1982 [4], which has beenused in vast area of computer science such as data mining,pattern recognition, machine learning and knowledgeacquisition etc [5].One of the first methods that can be used for servicediscovery is UDDI which is used for web service publicationand discovery. The current web service discovery mechanismis based on the standard of UDDI [6]. In UDDI, XML is usedto describe data in business services. UDDI process searchesqueries according to keywords and classification information.There is limitation with the discovery mechanism of UDDI.Firstly, machine can read XML data, but it can not understandXML data. Different query keywords may be semanticallyequivalent, whereas UDDI can not infer any information fromkeywords or tModels it can easily make mistake. Secondly,search by keywords and taxonomy is not suitable for webservice discovery. Furthermore, UDDI does not support searchby service capabilities and other properties [7]. This makesUDDI search method a low precision method [6].By advent of semantic web, services can be annotated withmetadata for enhancement of service discovery. One of theearliest to add semantic information is DAML-S [8]. DAML-Suses semantic information for discovering Web services.DAML-S uses ontological description to express web servicecapacity and character.
165http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 5, August 2010
OWL-S is an OWL [9] based ontology for encodingproperties of Web services. OWL-S technology is used tofacilitate service annotation and matching. OWL-S ontologydefines a service profile for encoding a service description, aservice model for specifying the behavior of a service, andservice grounding for how to invoke the service. Actually, byusing domain ontology described in OWL, using specialsoftware such as protégé [10], a service discovery processinvolves a matching between the profile of a serviceadvertisement and the profile of a service request. The serviceprofile describes the functional properties such as inputs,outputs, preconditions, and effects, and non functionalproperties such as service name, service category, and aspectsrelated to the quality of service.In [11] a quantification standard for semantic servicematching has been presented that modifies the classicalmatching algorithm based on OWL-S. Matching algorithm hasused the quantification standard of service matching and OWL-WS. In [12] service composition algorithm has constructed amathematical model and converted it to the shortest pathproblem in order to find process that can satisfy customer needin best conditions.In [7] an approach has been developed for integratingsemantic features into UDDI. The approach uses a semanticmatchmaker that imports OWL-S based semantic markups andservice properties into UDDI. The combination of OWL-S andUDDI shows there could be a service discovery which supportsweb service expression while UDDI is used. The matchmaker,therefore, enables UDDI to store semantic information of webservices and process service search queries based on semanticsimilarity of web service properties [7].The above-mentioned methods facilitate service discoveryin some way. However, when matching service advertisementswith service requests, these methods assume that serviceadvertisements and service requests use consistent properties todescribe relevant services. But for a system such as Grid with alarge number of resources and users which have their ownpredefined properties to describe services, it can't be true thatservice advertisements and service requests use consistentproperties to describe services. In other words, some propertiesmay be used in service advertisement that may not be used byservice request. So, an approach must be taken intoconsideration to deal with uncertainty of service propertieswhen matching service advertisements with service requests.Rough set theory is a new mathematical theory which dealswith uncertainty and vagueness [13].By moving toward the age of information, a hypothesis canformulate the human knowledge in the systematic form, andintroduce an approximate description that is reliable andanalyzable. This important subject is applicable by a fuzzysystem [14]. In our previous works [15, 16], we have usedclassic and dynamic rough set theory to deal with uncertaintyand vagueness. Using fuzzy system in matchmaking phasemake it possible to find resources met the requested propertieswith more precision and speed. Matchmaking phase is done byusing fuzzy system expressed according to properties of registered resources and user preferences which assigned tothese properties.The remainder of this paper is organized as fallows. Part IIis a review of related works. Part III is a description of theclassical rough set theory, part IV is a description of thedynamic rough set theory. Part V is the description of algorithm implemented and used in this paper, part VI is acomparison of our algorithm with dynamic and classical roughset based matchmaking algorithms [15,16], and finally part VIIis the conclusion and future works.II.
 
R
ELATED
W
ORKS
 While the grid environment moves towards a service-oriented computing infrastructure, service discovery isbecoming a vital part of this environment. One of the earliestmethods for service publication and discovery is UDDI whichonly supports keyword matches and does not support anysemantic service. DAML-S is the earliest to add semanticinformation for discovering web services [18]. DAML-S offersenough semantic information expressing Web service capacityand character with ontological description of web services. Inpast few years, a great amount of studies have been carried outon the basis of OWL-S, such as semantic expression servicebundling [19], ontology-based service matching [19], OWL-Sand UDDI combination [17]. In the [21] a metric is proposed tomeasure the similarity of semantic services annotated withOWL ontology. Similarity is calculated by defining theintrinsic information value of a service description based on theinferencibility of each of OWL constructs. All the abovemethods do not support uncertainty in properties. Rough settheory is used for dealing with vagueness and missing data inlarge variety of domains. So, compared with the work mentioned above, rough set theory can tolerate uncertainproperties in matching resources. In the [16] classical rough setbased algorithm has been proposed to deal with uncertainty andvagueness. Whereas grid environment is dynamic, usingclassical (static) rough set theory can not seem proper to use[15]. In this paper, our algorithm works in two steps. The Firststep is candidate optimization component which optimize thecandidate set using dynamic rough set theory. The Second stepis fuzzy matchmaking component which ranks resources usingfuzzy system according to requested resource.III.
 
C
LASSICAL
R
OUGH
S
ET
T
HEORY
Rough set theory which is proposed by Pawlak, in 1982,has been proved to be a good mathematical tool to describe andmodel uncertainty and imprecision. It has been widely appliedin artificial intelligent, pattern recognition, data mining, faultdiagnostics etc [21]. There are many advantages of rough setstheory; for example, no preliminary or additional information isneeded and only the facts in the data are considered.Let:
 
U: a set of N registered resources, U= {u
1
, u
2
, …, u
N
},N
1.
 
P: a set of M properties used to describe the N registeredresources of the set U, P = {p
1
, p
2
, …, p
M
} , M
2.
166http://sites.google.com/site/ijcsis/ISSN 1947-5500
 
(IJCSIS) International Journal of Computer Science and Information Security,Vol. 8, No. 5, August 2010
 
Q: a set of K registered resource properties relevant to aresource request R in terms of resource ontology
whose
 irrelevant properties have been removed,Q = {q
1
, q
2
, …, q
K
} , K
1, and Q is a subset of P.
 
R: a set of L requested resource properties with theirweights,R={(r
1
,w
1
), (r
2
,w
2
), …, (r
L
,w
L
)}, L
1.
According to the rough set theory for a given set X thereare:
[ ]
}{
 X  X  x X Q
Q
=
(1)
[ ]
}{
φ 
=
 X  X  x X Q
Q
(2)
In which
 X Q
is the lower approximation and
 X Q
is theupper approximation of X in terms of properties set Q. X is asubset of U and Q is a sub set of P.
 X 
 
PQ
 
So for a property
Qq
, we can say that:
 
,
 X Q x
x definitely is a member of X and definitelyhas property q.
 
 X Q x
, x probably is member of X and probably hasproperty q.
 X  x
, x absolutely is not a member of X andabsolutely does not have property q.The Most important part of rough set theory is attributereduction. Some attributes are dependent on other attributes inattributes set, so they are not necessary to be considered inmatching phase. According to rough set theory we are:
 D X 
 X  DPOS
 / 
)(
=
(3)
α 
=
 X  D
=
),(
γ  
(4)
In Which C and D are subsets of property set P. as shown in[13], D totally depends on C if 
α 
=1 Or D partially (in adegree of 
α 
) depends on C if 
1
<
α 
.Since existing works need to find exact match betweenrequested resources and registered resources, it is difficult tofind exact matching. So by using rough set theory, the need of exact match has been removed
.
IV.
 
D
YNAMIC
R
OUGH SET
T
HEORY
Although rough set theory is being used in various rangesof research such as data mining, pattern recognition, decisionmaking and expert system, it is suitable for static knowledgeand data. In fact, in a classical rough set theory, subset X of universal set U is a static set without considering the dynamicproperties it can have. In the real word, most informationsystems have dynamic properties so that the rate of participantand disappearance of entities in these systems is high. WhereasPawlak’s rough set theory can only deal with static informationsystem, using a dynamic method to deal with uncertainty andprocess information system will have more efficiency.By using dynamic rough set theory, considering dynamicproperties of an information system will be possible. Dynamicrough set theory uses outward and inward transfer parametersto expand or contract X set in classical rough set.According to [22], dynamic rough set theory has beendefined as follows:Suppose A= (U, P) is an information system,
P
and
 X 
. For any
 x
, we have:
Xxas, ][][)(
),(
=
 X 
 x X  x x
 ρ 
(5)
X~xas, ][][1)(
),(
=
+
 X 
 x X  x x
 ρ 
(6)
)(
),(
 x
 X 
 ρ 
is called outward transfer coefficient and
)(
),(
 x
 X 
+
 ρ 
is called inward transfer coefficient of element x aboutT. In real computation, outward and inward transfercoefficients are been choose as constant amounts. In fact
]1,0[)(
 X 
and
]1,0[)(
+
 X 
are outward transferstandard and inward transfer standard of elements of X aboutT, respectively.Inflated dynamic main set of X is defined as below:
}.1)(),(~{)(
),(
<=
+++
 X 
 X  X  x x X  M 
ρ 
(7)
And inflated dynamic assistant set is defined as:
)}.(0),(~{)(
),(
 X  X  x x X  A
 X 
+++
<=
ρ 
(8)
+
 X 
is
called inflated dynamic set of X about T and definedas:
 
).(
 X  M  X  X 
++
=
(9)
The formulas (5-9) show that we can expand X accordingto T. we can also contract X according to T. for this reason wehave:
}.1)()(,{)(
),(
<=
 X  X  X  x x X  M 
 X 
ρ 
(10)
In which
 
)(
 X  M 
is defined as contracted dynamic set of Xabout T and also contracted dynamic assistant set is defined as:
)}.()(0,{)(
),(
 X  X  X  x x X  A
 X 
<=
ρ 
(11)
And
 
 X 
called contracted dynamic set is defined as:
.
=
 M  X  X 
(12)
167http://sites.google.com/site/ijcsis/ISSN 1947-5500

You're Reading a Free Preview

Download
/*********** DO NOT ALTER ANYTHING BELOW THIS LINE ! ************/ var s_code=s.t();if(s_code)document.write(s_code)//-->