Simplified Multi-objective Optimization for Flexible IoT Edge Computing

Abstract

In large-scale, cloud-based IoT systems, challenges such as increased network load, response delays, and invasion of privacy are serious concerns in recent years. As a solution, edge computing (EC) has been introduced into these IoT systems. However, if too much of the cloud functionality is migrated to the edge, then the collected data cannot be shared between IoT systems, thereby decreasing its usefulness. We propose a multi-agent-based flexible IoT-EC architecture that balances the global optimization by the cloud with local optimization at the edges to optimize the roles of the cloud and edge servers dynamically. In this paper, we explain the simplified multi-objective optimization solutions for IoT EC. We also execute a simple traffic simulation to evaluate the proposed method.