Mobile Networks Accessibility and Retainability States Prediction Using Markov Chain

dc.contributor.advisorDereje, Hailemariam (PhD)
dc.contributor.authorTesfaye, Addisie
dc.date.accessioned2022-02-07T09:33:16Z
dc.date.accessioned2023-11-04T15:13:02Z
dc.date.available2022-02-07T09:33:16Z
dc.date.available2023-11-04T15:13:02Z
dc.date.issued2021-11
dc.description.abstractTo deliver reliable services to customers mobile network operators monitor their network performance using different key performance indicators (KPIs). Network accessibility and retainability are two important KPIs that need to be assessed and monitored on daily basis by mobile operators. Monitoring KPIs is a time consuming and laborious task. Moreover, taking corrective action based on monitored KPIs is a reactive approach and contributes to network and service degradation until corrective action is taken. Historical network performance data available with operators contains not only information about network status at the measurement instant but can be analyzed to provide information about possible trends and patters about the network. So, a systematic way of monitoring and predicting the network performance status is required to guaranty the service quality. In this thesis, a Markov chain (MC) model is used to predict, in a probabilistic sense, the accessibility and retainability states of a mobile network. Four months operator data is collected from 1,530 Universal Mobile Telecommunication System (UMTS) cells to formulate the MC model and investigate the prediction accuracy. Two scenarios are considered: first, two independent four-state MC models are used to separately model the accessibility and retainability state of a cell. Second, both accessibility and retainability states of the cell are jointly modeled using a single MC model with sixteen states. The proposed predictive model for the first scenario achieves an accuracy of 96.09% when predicting accessibility and 96.87% when predicting retainability separately, while 94.61% prediction accuracy is obtained when predicting accessibility and retainability jointly as in the second scenario. The result indicates that it is possible to monitor and predict the performance of a network using the joint modeling approach instead of using separate models. In addition, the sixteen states help to see the probability of occurrence of both accessibility and retainability at the same time; for instaen_US
dc.identifier.urihttp://etd.aau.edu.et/handle/123456789/29944
dc.language.isoen_USen_US
dc.publisherAddis Ababa Universityen_US
dc.subjectAccessibilityen_US
dc.subjectRetainabilityen_US
dc.subjectKPIen_US
dc.subjectUMTSen_US
dc.subjectMarkov chainen_US
dc.titleMobile Networks Accessibility and Retainability States Prediction Using Markov Chainen_US
dc.typeThesisen_US

Files

Original bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
Tesfaye Addisie.pdf
Size:
1.91 MB
Format:
Adobe Portable Document Format
License bundle
Now showing 1 - 1 of 1
No Thumbnail Available
Name:
license.txt
Size:
1.71 KB
Format:
Plain Text
Description: