Optimization of Round-Robin Arbitration for Low Power

Abstract

This paper proposes an energy-efficient task scheduling strategy by optimizing the Round-Robin (RR) arbitration mechanism using a hybrid Artificial Bee Colony with Whale Optimization Algorithm (ABC & WOA) in conjunction with Dynamic Voltage Scaling (DVS). While RR scheduling is widely used for its simplicity and fairness, it typically lacks energy awareness—an increasingly critical concern in energy-constrained and battery-powered embedded systems. This study addresses this gap by incorporating DVS and optimized idle-time sleep states to significantly reduce both total energy consumption and average energy per task. The proposed hybrid ABC-WOA algorithm demonstrates superior performance over its individual counterparts (ABC and WOA), offering improvements in energy efficiency, convergence speed, task completion rate, and execution time. The experimental results validate that the integrated DVS and ABC-WOA approach is a robust and scalable solution for low-power task scheduling, particularly in embedded systems and real-time environments.