HARQ Control Scheme by Fuzzy Q-Learning for HSPA+

In this paper, we propose a hybrid automatic repeat request (HARQ) control scheme by using fuzzy Q-learning algorithm for high speed packet access evolution (HSPA(+)) systems. The challenge of the HARQ control in HSPA(+) systems is the selection of MIMO (multiple-input multiple-output) configuration mode and MCS (modulation and coding scheme) level for new data transmission of HARQ, under the situation that the channel quality indication has report delay. The fuzzy Q-learning-based HARQ (FQL-HARQ) scheme can determine an appropriate MIMO configuration mode and MCS level for new data transmission so as to maximize the system throughput under the guaranteed block error rate (BLER) requirement. Simulation results show that, the proposed FQL-HARQ scheme increases the system throughput by up to 70.09% as compared to the conventional adaptive threshold method [4].