RATSS: A RISK-AWARE MULTI-OBJECTIVE TASK SCHEDULING STRATEGY FOR OPTIMIZED PERFORMANCE IN CLOUD-EDGE ENVIRONMENTS
TL;DR
A Risk-Aware Task Scheduling Strategy (RATSS) that incorporates variability into the optimization process by minimizing the standard deviation of makespan and waiting time, thereby reducing execution uncertainty and demonstrating improved stability, fairness, and overall system performance.
Abstract
Efficient task scheduling in cloud-edge environments remains a significant challenge due to resource heterogeneity, varying workloads, and conflicting objectives of Cloud Service Providers (CSPs) and Cloud Service Users (CSUs). Existing multi-objective approaches primarily optimize average-based metrics such as makespan and waiting time while neglecting performance variability, leading to workload imbalance and unstable execution. To address this limitation, this study proposes a Risk-Aware Task Scheduling Strategy (RATSS) that incorporates variability into the optimization process by minimizing the standard deviation of makespan and waiting time, thereby reducing execution uncertainty. Reducing makespan variability improves workload consistency and resource utilization for CSPs, while minimizing waiting-time variability ensures fairness and reduced delays for CSUs. A composite metric, Satisfaction Ratio (SR), is introduced to jointly capture stability and fairness. The proposed approach is evaluated using the NASA iPSC workload dataset and compared with Non-dominated Sorting Genetic Algorithm (NSGA)-I, NSGA-II, Multi-Objective Particle Swarm Optimization, and the Strength Pareto Evolutionary Algorithm II. Results show average reductions of 48.05% and 63.23% in makespan and waiting-time variability, respectively, demonstrating improved stability, fairness, and overall system performance.