Efficient Resource Allocation Framework for Network Slicing-enabled Open RAN

The transition to 5G technology has revolutionized cellular networks, leading to an unprecedented demand for high-bandwidth and low-latency services. To meet these evolving demands, next-generation mobile networks are turning to innovative solutions such as network slicing within the open radio access network (RAN) architecture. This paper focuses on addressing the challenges associated with dynamic virtual network function (VNF) mapping and resource allocation in terms of both communication and computing resources. The ultimate goal is to simultaneously maximize the total reward for admitted slices and minimize the associated cost by allocating resources to the Open RAN slices catering to diverse service types. To efficiently solve the formulated problem, we decompose it into two sub-problems tailored to different timescales which are executed at the different layers within the Open RAN architecture. To solve the non-convex long-term problem, we first transform it to a more tractable form and employ the successive convex approximation (SCA) method to achieve at least a locally optimal solution. Numerical simulations demonstrate the efficacy and quick convergence of the proposed algorithm, highlighting its potential to improve network slicing performance and support diverse applications in future wireless networks.