Descripció del projecte
The project entitled “Cloud Continuum Foundational Models Through Reinforcement Learning Strategies” aims to develop foundational models capable of representing and optimizing the deployment of applications in the cloud continuum, an ecosystem that ranges from centralized cloud infrastructures to distributed environments such as edge and fog computing. Using advanced reinforcement learning techniques, intelligent agents will be trained to make optimal decisions in real time, improving resource allocation, reducing latency and increasing operational efficiency. From these agents, models of the world that capture the complex dynamics of the continuum will be extracted, which will serve as a basis for building generalizable foundational models.
These models will enable the creation of digital twins capable of simulating real-world scenarios, anticipating failures, assessing the impact of deployments, and facilitating interoperability between multiple operators. The project will also explore approaches such as federated learning and multi-agent reinforcement learning to address infrastructure heterogeneity and fragmentation, thus contributing to the advancement of intelligent, resilient, and scalable systems in the next-generation cloud computing arena.
