Page - 359 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

In theworkproposed in[13], the author has addedanewcomponent calledOoSBF (QualityofServiceBootstrappingFramework)whichconsiders theneed forQoSquali- fication integrationwhenpublishinganewServiceandbefore it isused. QoS is an important factor for the evaluation of distributed paradigms, indeed sev- eral studiesmention that the justification ofQoS is done at the time of the selection of serviceorduringexecution.Several quality services arementioned in thiswork,wecan cite:Latency,Execution time,Response time,Throughput,Availability,Reliability, and Accessibility. Intheworkproposedin[16], theauthoraddedtotheUDDIregisteranewcomponent calledMEC,which deals with the evaluation of possible components in order to help the applicant of the service inquestion to choose theWebservice to invoke. It confirms that, in order to carry out a better selection ofweb services, itwas necessary to use the different non-functional criteria instead of the functional criteria since in general they are the same functionality and since this is not a single objective solution, the author proposedamulti-criteria evaluation. In [17],TOPSIS is amulti-criteriamethod thatmakes it possible to classify a set of alternatives inorderofchoiceaccording toseveralcriteria.Thebasic ideaof thismethod is to choose a solution that comes closest to the ideal solution (better on all the criteria) andtomoveawayfromtheworst solutionasmuchaspossible.However thismethodhas aweaknesswhere it doesnotdetermine theweights to theevaluationcriteria. In [18], we have proposed an approach based onTOPSISmethod forweb service selection. In this work, we did not take into consideration the weights relating to the criteria, theseweights arefilledmanuallyby thedesigner. 3. Proposedapproach In this section,wepresent ourproposedapproach that describes the ideaof the solution thenwefocuson theagent-basedsolution. Thisapproach in theFigure1 isacombinationof twomethodsAHPandTOPSISto serve theclients thebest service. Ouragentbasedsolution isdepicted in theFigure2. In thisfigure,weshowthe life cycle of a user request for a givenWeb service. The request is a keywordwith a set of criteria.First, inourapproach theproviderpublishesasetofWebservices in theservice registry(UDDI).ItassociateswitheachservicedescriptionasetofQoSinformationsuch asavailability, reliability,etc. (A.1 in thefigure).TheUDDIsaves theQoSinalocaldata base(seeA.2).Second, theRequester Agent receives requests fromtheserviceclient. Hence, theRequester Agent sends theWeb service requestswith the desiredQoS to Broker Agent (B.1 in the Figure). The latter, forwards these requests to the service registry in order to find all the services that answer to the request and the desiredQoS (seeB.2).Once theBroker Agent receives the list of similar services (B.3 andB.4 in theFigure), it interactswithWeighting Agent inorder tocalculate theweights thatare relatedtothedesiredQoScriteria(seeB.5). In thisstep,weuseAHPmethodtocalculate suchweights.Theobtained list ofweighted services is sent to theRanking Agent (see B.6). The Ranking Agent classifies this list ofweighted services by assigning scores to each one.We use for this, the TOPSISmethod. In this way, we can obtain the best servicebasedontheassignedscores.Finally, theRanking Agentsends thebest service to theRequester Agentwhich is ananswer to the serviceclient request. H.Belouaaretal. /AHPandTOPSISMethods forWebServiceSelection 359