Federate Learning Training set up

    The increasing energy demand in data centers poses a severe operational challenge to telecommunication operators. This challenge arises from the growing volume of data traffic generated by diverse user demands, which consequently escalates energy consumption and operational costs. To alleviate this burden, the present work proposes an adaptive temperature regulation framework that leverages a physics-informed DeepONet to model heat generation dynamics within the data center. The DeepONet model provides accurate thermal insights to the HVAC system, enabling intelligent decision-making for maintaining optimal temperature levels based on real-time thermal conditions.     The overall control problem is formulated as a Markov Decision Process (MDP) and solved using the Asynchronous Advantage Actor\u2013Critic (A3C) algorithm to optimize both energy consumption and thermal stability. Simulation results demonstrate that the proposed DeepONet\u2013A3C-HVAC framework achieves high prediction accuracy and significantly reduces energy expenditure.