Comparing time. A New Approach To The Problem Of Time Synchronization In a Multi-agent System

Abstract

A live time-dependent tasks execution based on a distributed multi-agent network relies strongly on the time synchronization ability of the system. Desynchronisation of the system may lead to incorrect results or even failure at performing a certain tasks. This paper presents a novel synchronization algorithm that addresses the challenge of time discrepancies. It employs a unique approach of dynamically accelerating and decelerating time. This method combines the strengths of consensus algorithm and a system based on Nesterov's accelerated gradient descent. The result is a real-time synchronization solution that avoids the common issue of dropped time points encountered in traditional algorithms.

Keywords

• multi-agent systems
• time synchronization
• distributed networks
• distributed time synchronization
• distributed optimization