A Network-Assisted Approach for RAT Selection in Heterogeneous Cellular Networks

摘要: When several radio access technologies (e.g., HSPA, LTE, WiFi, and WiMAX) cover the same region, deciding to which one mobiles connect is known as Radio Access Technology (RAT) selection problem. To reduce network signaling processing load, decisions are generally delegated mobile users. Mobile users aim selfishly maximize their utility. However, they do not cooperate, may lead performance inefficiency. In this paper, overcome limitation, we propose a network-assisted approach. The provides information for make more accurate decisions. By appropriately tuning information, user globally expected meet operator objectives, avoiding undesirable states. Deriving formulated semi-Markov decision process (SMDP), optimal policies computed using Policy Iteration algorithm. Also, since parameters be easily obtained, reinforcement learning approach introduced derive what signal mobiles. performances of optimal, learning-based, heuristic policies, such blocking probability average throughput, analyzed. thresholds pertinently set, our achieves very close solution. Moreover, although it lower performance, learning-based algorithm has crucial advantage requiring no prior parameterization.