A Reinforcement Learning Integrating Distributed Caches for Contextual Road Navigation

Literature Review

Reinforcement learning is a paradigm useful to yield the problem of autonomous learning in dynamic environments, assuming for instance no prior information about the actions to perform (Stafylopatis and Blekas, 1998; Veres and Moussa, 2019). Compared to the model-based approach specifying an ‘a priori’ time-dependent behavior for a real time system, e.g., (Boukharrou et al, 2017), it appears to be more adequate to capture dynamicity on-the-fly, handling immediate adaptations of the execution parameters.