Remote Electrical Tilt (RET) optimization is an efficient method for adjusting the vertical tilt angle of Base Stations (BSs) antennas in order to optimize Key Performance Indicators (KPIs) network. Reinforcement Learning (RL) provides a powerful framework RET because its self-learning capabilities and adaptivity environmental changes. However, RL agent may execute unsafe actions during course interaction, i.e., resulting undesired network performance degradation. Since reliability services...

6G will move mobile networks towards increasing levels of complexity. To deal with this complexity, optimization network parameters is key to ensure high performance and timely adaptivity dynamic environments. The the antenna tilt provides a practical cost-efficient method improve coverage capacity in network. Previous methods based on Reinforcement Learning (RL) have shown effectiveness for by learning adaptive policies outperforming traditional methods. However, most existing RL are...

We address the problem of Remote Electrical Tilt (RET) optimization using off-policy Contextual Multi-Armed-Bandit (CMAB) techniques. The goal in RET is to control orientation vertical tilt angle antenna optimize Key Performance Indicators (KPIs) representing Quality Service (QoS) perceived by users cellular networks. Learning an improved update policy hard. On one hand, coming up with a new online manner real network requires exploring updates that have never been used before, and...

Controlling antenna tilts in cellular networks is imperative to reach an efficient trade-off between network coverage and capacity. In this paper, we devise algorithms learning optimal tilt control policies from existing data (in the so-called passive setting) or actively generated by (the active setting). We formalize design of such as a Best Policy Identification (BPI) problem Contextual Linear Multi-Arm Bandits (CL-MAB). An arm represents update; context captures current conditions;...

Safe interaction with the environment is one of most challenging aspects Reinforcement Learning (RL) when applied to real-world problems. This particularly important unsafe actions have a high or irreversible negative impact on environment. In context network management operations, Remote Electrical Tilt (RET) optimisation safety-critical application in which exploratory modifications antenna tilt angles base stations can cause significant performance degradation network. this paper, we...

We investigate the problem of Remote Electrical Tilt (RET) optimization using off-policy learning techniques devised for Contextual Bandits (CBs). The goal in RET is to control vertical tilt angle antennas at base stations optimize key performance indicators representing Quality Service (QoS) perceived by users cellular networks. Learning an improved update policy hard. On one hand, coming up with a online manner real network requires exploring updates that have never been used before, and...

Using analog beamforming in mmWave frequency bands we can focus the energy towards a receiver to achieve high throughput. However, this requires network quickly find best downlink beam configuration face of non-IID data. We propose personalized Federated Learning (FL) method address challenge, where learn mapping between uplink Sub-6GHz channel estimates and heterogeneous scenarios with characteristics. also devise FedLion, FL implementation Lion optimization algorithm. Our approach reduces...

Within the business context, study of innovator’s personal traits is a fundamental analysis to be carried out in order understand complete way howthe innovation process takes place and find which are necessary make ithappen. Based on five-factor model, this work explores how personality arerelated successful ways innovate provides contemporary example analysisby studying one nowadays most innovative personalities: Elon Musk. By comparing model with concrete experimental case, conclusions...

Remote Electrical Tilt (RET) optimization is an efficient method for adjusting the vertical tilt angle of Base Stations (BSs) antennas in order to optimize Key Performance Indicators (KPIs) network. Reinforcement Learning (RL) provides a powerful framework RET because its self-learning capabilities and adaptivity environmental changes. However, RL agent may execute unsafe actions during course interaction, i.e., resulting undesired network performance degradation. Since reliability services...

Using analog beamforming in mmWave frequency bands we can focus the energy towards a receiver to achieve high throughput. However, this requires network quickly find best downlink beam configuration face of non-IID data. We propose personalized Federated Learning (FL) method address challenge, where learn mapping between uplink Sub-6GHz channel estimates and heterogeneous scenarios with characteristics. also devise FedLion, FL implementation Lion optimization algorithm. Our approach reduces...

Safe interaction with the environment is one of most challenging aspects Reinforcement Learning (RL) when applied to real-world problems. This particularly important unsafe actions have a high or irreversible negative impact on environment. In context network management operations, Remote Electrical Tilt (RET) optimisation safety-critical application in which exploratory modifications antenna tilt angles base stations can cause significant performance degradation network. this paper, we...

We address the problem of Remote Electrical Tilt (RET) optimization using off-policy Contextual Multi-Armed-Bandit (CMAB) techniques. The goal in RET is to control orientation vertical tilt angle antenna optimize Key Performance Indicators (KPIs) representing Quality Service (QoS) perceived by users cellular networks. Learning an improved update policy hard. On one hand, coming up with a new online manner real network requires exploring updates that have never been used before, and...

Controlling antenna tilts in cellular networks is imperative to reach an efficient trade-off between network coverage and capacity. In this paper, we devise algorithms learning optimal tilt control policies from existing data (in the so-called passive setting) or actively generated by (the active setting). We formalize design of such as a Best Policy Identification (BPI) problem Contextual Linear Multi-Arm Bandits (CL-MAB). An arm represents update; context captures current conditions;...

6G will move mobile networks towards increasing levels of complexity. To deal with this complexity, optimization network parameters is key to ensure high performance and timely adaptivity dynamic environments. The the antenna tilt provides a practical cost-efficient method improve coverage capacity in network. Previous methods based on Reinforcement Learning (RL) have shown great promise for by learning adaptive policies outperforming traditional methods. However, most existing RL are...