A Dual Phase Genetic Algorithm with Aggregated Search for Fast Initial Access in 5G Millimeter Wave Communication

5G Millimeter Wave (mmWave) Communication between the base station (BS) and the user equipment (UE) involves a Multiple-Input Multiple-Output (MIMO) system where both BS and UE have many antennas. Initial Access (IA) in this context is the problem of establishing a directional link between the BS and UE, but finding the optimal beams can be prohibitively expensive in terms of delay and computation. Genetic Algorithms (GAs) can solve complex problems effectively, and in this case, they can be used to iteratively search for the optimal beams. We propose a dual phase GA that splits the GA process into two successive phases that uses different operations in each phase. It also navigates the search space in a smart manner, increasing the convergence rate to the optimal beamformer per iteration. We have analyzed the effect of this approach in terms of Capacity achieved vs number of transmit and receive antennas at BS and UE, total transmitted power, and number of iterations. It shows improved performance in terms of maximum Capacity achieved, reduced power consumption, and especially reduced IA delay.