Qos Aware Adaptive Data Dissemination In Mobile Edge Computing Ecosystem

Abstract

In the evolving paradigm of Mobile Edge Computing (MEC), ensuring Quality of newlineService (QoS) in dynamic, real-time environments present significant challenges due newlineto fluctuating network topologies, heterogeneous resource constraints, and increasing newlinedemand from data-intensive applications. This thesis work addresses these challenges newlineby implementing a series of adaptive, intelligent, and QoS-aware models rooted in newlinebioinspired and machine learning approaches, aimed at enhancing traffic control, data newlinedissemination, and resource scheduling in MEC deployments. A unique Dynamic newlineTraffic Flow Control (DTFC) framework, combined with a QoS-aware Adaptive Data newlineDissemination Engine (QADE), was presented to address the issues of network newlinecongestion and delay. Based on temporal and geographical parameters, this model newlineadaptively manages communication flows by utilizing a hybrid Elephant Herding newlineParticle Swarm Optimizer (EHPSO) in conjunction with reinforcement learning newlineapproaches. During extensive simulations, the system showed notable gains in newlinelatency, throughput, energy efficiency, and packet delivery ratio. Additionally, using newlineFlower Pollination Optimization (FPO) and the predictive ability of a VARMAx newline(Vector Autoregressive Moving Average with exogenous variables) model, a newlinebioinspired resource scheduling model was created. By taking into account a wide newlinerange of task and resource characteristics, our hybrid architecture effectively mapped newlinetasks to virtual machines. Additionally, it enabled the dynamic recalibration of virtual newlinemachine capacity by predicting future workloads, thereby improving scheduling newlineeffectiveness, energy conservation, and deadline adherence. Extensive tests on real- newlineworld datasets confirmed that the suggested models performed well in comparison to newlineexisting techniques. newline