Genetic Algorithm in Resource Allocation of RAN Slicing with QoS Isolation and Fairness

Radio access network (RAN) slicing is a challenging task, as different levels of isolation and fairness requirements need to be considered. In the context of resource allocation for RAN slicing, this paper addresses three major issues including maximizing the number of qualified users who are assigned the required bitrates, reducing the inter-slice QoS isolation between neighbouring cells caused by inter-cell interference and maximizing the fairness among all the slices. The proposed real number encoded Genetic Algorithm requires no prior knowledge about the network traffic models, provides high flexibility of slicing scheme, converges very quickly to evolve optimal/suboptimal solutions, and the dynamic weights of different objectives can adapt to the changes of users' scenarios.