Energy Saving and Interference Coordination in HetNets Using Dynamic Programming and CEC

摘要: Energy efficiency in cellular networks has gained great relevance due to the increasing power supply demands new generation heterogeneous network (HetNet). On other hand, interference coordination is a key aspect HetNets resource management which directly affects performance and energy consumption. Although these two are intertwined, they have been studied separately so far. In this paper, we address joint problem of saving HetNets. We formulate as finite horizon Markov decision process (MDP) leveraging facts: user traffic usually follow periodic patterns, knowledge prediction load crucial order select efficient configurations. Quality Service (QoS) defined ratio users meeting requirement specified by operator, allows us account for minimum QoS including constraint formulation MDP. To MDP, propose an approximate dynamic programming (ADP) algorithm selects farsighted configurations with guarantees achieving near optimal performance. This ADP built upon: 1) certainty equivalent control principle, simplifies complexity 2) machine learning techniques: neural polynomial regressor allow predict consumption advance. evaluate our proposal LTE-A simulator following 3GPP guidelines, results obtained show that mechanisms notably improvement compared disjoint control, terms both savings guarantees. Moreover, advantage being adaptable operator requirements.