Online Learning for Rate-Adaptive Task Offloading Under Latency Constraints in Serverless Edge Computing

We consider the interplay between latency constrained applications and function-level resource management in a serverless edge computing environment. develop game theoretic model of interaction rate adaptive load balancing operator under function-oriented pay-as-you-go pricing model. show that perfect information, strategic can be formulated as generalized Nash equilibrium problem, use variational inequality theory to prove admits an equilibrium. For case imperfect we propose online learning algorithm for maximize their utility through adaptation reservation. proposed converge equilibria achieves zero regret asymptotically, our simulation results good system performance at equilibrium, ensures fast convergence, enables meet constraints.