Improved genetic algorithm based intelligent resource allocation in 5G Ultra Dense networks

As a key technology, it is expected that the Ultra Dense network (UDN) architecture will play a key role in supporting the fifth generation (5G) of mobile communication technologies, especially for hotspot and blind wireless areas. Energy Efficiency (EE) and Spectrum Efficiency (SE) are two important metrics in the 5G UDN. Generally, they can not obtain optimal results simultaneously. To balance the tradeoff of them, in this paper a multi-objective optimization problem (MOOP) is formulated and an improved version of nondominated sorting genetic algorithm II (NSGA-II) based intelligent approach is proposed which enables small cell users to optimize their downlink performance of EE and SE by jointly allocating transmission power and resource blocks. Simulation results show that the proposed algorithm yields significant performance gains when compared with existing exhaustive search and weighted sum method. Furthermore, the convergence and computational complexity of our proposed algorithm are studied.