Page - 364 - in Intelligent Environments 2019 - Workshop Proceedings of the 15th International Conference on Intelligent Environments

⎛ ⎜⎜⎜⎜⎝ Rel Th Cost Rt SMSWS 79,3 6,3 11 751 SMS 81 5,2 2 113,8 SendMessages 53,6 4,5 9 291,07 SendSMS 67 6,8 3 1308 SendSMSWorld 64,3 5,3 4 3103 ⎞ ⎟⎟⎟⎟⎠ The application of the AHPmethod gives for each criterion the weights that are depicted inTable4: Table4. CriteriaWeights Criteria Cost response time Reliability Throughput Weight 0,056 0,122 0,216 0,606 Theseweightsare the inputs for theRanking Agent thatuses theTOPSISmethod. TheTOPSISmethodstartbycreatingadecisionmatrix(presentedabove)whichcontains theWebservicesQoSvalues.Thismatrix isnormalizedas inTable5. Table5. Normalizedmatrix Services/Criteria Rel Th Cost Rt SMSWS 0,508 0,496 0,724 0,217 SmsWS 0,519 0,409 0,132 0,033 SendMessage 0,344 0,354 0,592 0,084 SendSMS 0,429 0,535 0,197 0,378 SendSMSWord 0,412 0,417 0,263 0,896 Thenormalizedmatrix isusedwithpreviousweights inorder tobuild theweighted normalizedmatrix.TheobtainedMatrix isdepicted inTable6. Table6. Weightednormalizedmatrix Alternative Rel Th Cost Rt SMSWS 0,110 0,301 0,041 0,026 SmSWS 0,112 0,248 0,007 0,004 SendMessage 0,074 0,215 0,033 0,010 SendSMS 0,093 0,324 0,011 0,046 SendSMSWord 0,089 0,253 0,015 0,109 Byapplyingthefinalstepinourapproach,weobtainthescoresforeachWebservice (seeTable 7).Thebest service is theonewith thehighest score. So inour example, the SendSMSis thebest recommendedservice to theclient. Table7. WebServiceRanking Alternative Score orderof choice SMSWs 0,72737 2 SmsWs 0,61391 3 SendMessage 0,45466 4 SendSMS 0,73874 1 SendSMSWord 0,27189 5 H.Belouaaretal. /AHPandTOPSISMethods forWebServiceSelection364