Machine intelligence in allocating bandwidth to achieve low-latency performance

Abstract

In this work, we present a complete rethink of the decision-making process in allocating bandwidth in a heterogeneous Fiber-Wireless network with machine intelligence. We highlight the use of an artificial neural network (ANN) at the central office to learn the uplink latency performance using multiple network and packet features. In turn, the trained ANN enables the central office to facilitate flexible bandwidth allocations under diverse network scenarios in meeting low-latency communication demands.