Cognitive radio with reinforcement learning applied to heterogeneous multicast terrestrial communication systems

This paper shows how channel assignment in heterogeneous multicast terrestrial communication systems can be improved using intelligence based on reinforcement learning. Two novel distributed channel assignment schemes with reinforcement learning applied are shown, which efficiently improves the speed and quality of channel assignment by limiting the reassignments, blocking and dropping rates. A weighting factor is used in this paper to determine the highest priority channels, and to help to control the performance of the system. It is found that reinforcement learning provides an efficient approach to reduce the needs of reassignments. At the same time, reassignment is a very good alternative to using blocking of new assignments to control dropping. Learning is categorized into 3 stages depending on the degree of effect it has on behavior.