Machine learning algorithms for inter-cell interference coordination

Abstract

The current LTE and LTE-A deployments require larger efforts to achieve the radio resource management. This, due to the increase of users and the constantly growing demand of services. For this reason, the automatic optimization is a key point to avoid issues such as the inter-cell interference. This paper presents several proposals of machine-learning algorithms focused on this automatic optimization problem. The research works seek that the cellular systems achieve their self-optimization, a key concept within the self-organized networks, where the main objective is to achieve that the networks to be capable to automatically respond to the particular needs in the dynamic network traffic scenarios.

Solution for Interference in Hotspot Scenarios Applying Q-Learning on FFR-Based ICIC Techniques

“…They have been widely discussed in the literature for their ability to adapt to changing scenarios and reduce site planning. For instance, the authors of [ 47 ] present various ML algorithms and how they have been employed to coordinate co-channel interference.…”

“…Among RL approaches, Q-Learning (QL) has been constantly discussed in the literature, especially because it does not require estimating the dynamics of the environment (it is model-free) [ 47 ].…”