MCDS: AI Augmented Workflow Scheduling in Mobile Edge Cloud Computing Systems

Workflow scheduling is a long-studied problem in parallel and distributed computing (PDC), aiming at efficiently utilizing compute resources to meet user's service requirements. Recently proposed methods leverage the low response times of edge platforms optimize application Quality Service (QoS). However, workflow applications mobile edge-cloud systems challenging due computational heterogeneity, changing latencies devices volatile nature workload resource To overcome these difficulties it important, but same time challenging, develop long-sighted optimization scheme with efficient modelling QoS objectives. In this work, we propose MCDS: Monte Carlo Learning using Deep Surrogate Models schedule systems. MCDS an Artificial Intelligence (AI) based approach that uses tree-based search strategy deep neural network surrogate model estimate long-term impact immediate actions for robust decisions. Experiments on physical simulated testbeds show can improve over state-of-the-art terms energy consumption, time, SLA violations cost by least 6.13, 4.56, 45.09 30.71 percent respectively.