Mobility robustness optimization and load balancing in self-organized cellular networks: Towards cognitive network management

Self-Organization networking (SON) consists of function sets which are responsible for automatically reliable configuring, planning and optimizing next generation mobile networks. Effective self-organization functions improve the level of network key performance indicators by determining optimal network setting and continuously finding efficient solutions that will be very hard for experts to distinguish. Most current self-organization networking functions apply rule-based recommended systems to control network resources in which performance metrics are evaluated and the effective actions are performed in accordance with a set of command sequences which such algorithms are too complicated to design, because rules and command sequences should be derived for each target index during each possible scenario. This research has proposed cognitive wireless networks as a fully intelligent approach to self-organization networking. We generalize the concept of network automation considering fuzzy-based self-organization networking functions as Q-learning problems in which, a framework is described to find the fuzzy optimal solution of linear programming optimization problem. The achieved results prove that the proposed cognitive approach, provides a prominent cellular framework for developing self-organization solutions, particularly where the relevance of metrics to the control indices is not clearly known. Also, assessment of the scheme in multiple-speed scenarios revealed that Q-learning load balancing obtains more accurate results compared to rule-based adaptive load balancing methods. This is particularly correct in dynamic networks, with high-speed users.