研究目的

To autonomously solve problems in the radio access network and to minimize their impact on the user through a root cause analysis system based on fuzzy logic and a genetic algorithm for learning the rule base.

研究成果

The proposed method of using a genetic algorithm to automatically learn the rule base of fuzzy controllers for mobile network self-diagnosis is robust and reduces operational expenditure while enhancing the quality of service. The system is adaptable to the reasoning of troubleshooting experts, easing knowledge acquisition and system output interpretation. The results are comparable or even better than those obtained when experts define the rules, with the benefit of not requiring direct expert intervention.

研究不足

The main limitation is the lack of existing databases of classified fault cases in which to test new diagnosis systems, particularly in LTE, which is starting to be deployed in current networks. Additionally, the algorithm's performance degrades with an increased number of PIs, requiring more complex individuals and longer execution times.