Development of Energy Efficient Algorithms for Edge Computing Based Artificial Intelligence Applications in Smart Cities

Abstract

The rapid expansion of smart cities demands the deployment of energy-efficient, intelligent systems at the network edge. Traditional cloud-centric artificial intelligence (AI) architectures are insufficient due to high latency, bandwidth constraints, and excessive energy consumption. In this paper, we explore the development of energy-efficient algorithms specifically designed for edge computing platforms supporting AI applications in smart cities. We first review the state of research, identify critical challenges, and propose a hybrid optimization framework combining lightweight neural networks and energy-aware task scheduling. Preliminary simulations demonstrate that our approach reduces energy consumption by up to 35% compared to conventional edge-AI methods, while maintaining near-optimal performance