Hierarchical SLA-aware multi-objective task scheduling in multi-access edge computing using fuzzy-AHP and queue stability modeling

Multi-Access Edge Computing (MEC) has emerged as an effective paradigm for supporting latency-sensitive and computation-intensive applications generated by mobile and IoT devices. However, efficient task scheduling in MEC environments remains challenging due to heterogeneous workloads, limited edge resources, and strict service-level agreement (SLA) requirements. This paper proposes a Hierarchical SLA-Aware Multi-Objective Edge Scheduling (HS-MOES) framework for efficient task allocation in MEC systems. The proposed approach integrates the Analytic Hierarchy Process (AHP) and fuzzy inference to perform multi-criteria task prioritization based on execution delay, energy consumption, and SLA violation risk. In addition, a queue-stability-aware scheduling mechanism is introduced to prevent congestion at edge servers, while an SLA-aware cloud migration strategy dynamically redirects tasks when edge resources become insufficient. The performance of the proposed framework is evaluated through large-scale simulations involving 50–200 mobile users and 5–30 edge virtual machines (VMs). HS-MOES is compared with conventional scheduling algorithms (FIFO, RR, SJF) and recent MEC scheduling methods including FAT-ETO, Fuzzy-DRL, and Enhanced Fuzzy Scheduling. Experimental results demonstrate that the proposed approach significantly improves system performance. In particular, HS-MOES reduces makespan by up to 24%, average task delay by 19%, and energy consumption per task by 17% compared with existing approaches, while maintaining higher resource utilization and lower SLA violation rates. Statistical analysis over 30 simulation runs with 95% confidence intervals confirms the robustness and stability of the proposed scheduling framework.