Page 379 - AI for Good Innovate for Impact

AI for Good Innovate for Impact

Once the long-term network configurations are optimized using GA, the twin configurations
are sent to the RAN Intelligent Controller (RIC) within the OpenRAN stack. The RIC then
communicates these optimal settings to the Distributed Unit (DU) and Radio Unit (RU), ensuring
that the network operates efficiently while meeting coverage and capacity requirements. 4.3 - 5G
2) Proximal Policy Optimization (PPO)

While Genetic Algorithms (GA) focus on global optimization and long-term planning, Proximal
Policy Optimization (PPO) is employed for real-time dynamic adjustments with the 5G base
station. PPO is a type of reinforcement learning (RL) that enables the network to adapt its
parameters in response to ongoing Key Performance Indicators (KPIs) such as signal strength
and energy consumption. Once the KPIs are maximized, the current configuration is sent as
feedback to GA for future changes in GIS. Hence PPO helps in building a closed loop system
for digital twin.

The functionalities of PPO are,

Adjusting Transmission Power and Beamforming: In real-time, PPO continuously monitors the
network conditions based on setup by GA and adjusts parameters like transmission power.
This ensures that signal strength is maintained across areas of high demand while minimizing
energy wastage in areas with lower user density or signal requirements.

Efficient Resource Allocation: PPO dynamically manages resource allocation, ensuring that
available spectrum and time slots are used efficiently. For example, it reduces transmission
power or offloads traffic during periods of low demand, thus optimizing energy consumption.

PPO operates within the RIC in the OpenRAN stack that communicates with the DU and RU in
real-time. The PPO adjusts network parameters based on live data and sends these updates
immediately to the OpenRAN components for ensuring energy efficiency. At the same time, it
sends parameters as a feedback loop to GA algorithm for future optimization considerations.

The proposed solution enables significant reductions in energy consumption and CO₂ emissions
by avoiding wasteful static configurations and overprovisioning. Moreover, its autonomous,
scalable nature positions it as a key enabler for green, sustainable 6G networks, supporting
global climate goals and sustainable smart city development.

Partners

Amrita University, ESRI ArcGIS, Indian Institute of Science, Indian Open Source Mobile Congress

2�2 Benefits of the use case

This use case strengthens digital infrastructure by enabling 5G networks to dynamically adapt
to environmental changes through the integration of AI technologies and ArcGIS geospatial
data. This enhances the responsiveness, efficiency, and sustainability of telecom systems.

By leveraging real-time ArcGIS data, the solution supports the seamless evolution of 5G
networks alongside urban growth. It helps reduce connectivity gaps and congestion, directly
benefiting smart city initiatives such as Internet of Things (IoT) deployments, autonomous
transportation systems, and emergency response networks.

343