Asymptotic behavior of the linear consensus model with delay and anticipation

TL;DR

This paper investigates how delay and anticipation influence the long-term behavior of linear consensus models, revealing that moderate anticipation promotes convergence while excessive anticipation can destabilize the system.

Contribution

It provides explicit conditions for asymptotic consensus in multi-agent systems considering delay and anticipation effects, applicable regardless of the number of agents.

Findings

Moderate anticipation generally promotes consensus.

High anticipation levels can destabilize the system.

Explicit parameter conditions for consensus are derived.

Abstract

We study asymptotic behavior of solutions of the first-order linear consensus model with delay and anticipation, which is a system of neutral delay differential equations. We consider both the transmission-type and reaction-type delay that are motivated by modeling inputs. Studying the simplified case of two agents, we show that, depending on the parameter regime, anticipation may have both a stabilizing and destabilizing effect on the solutions. In particular, we demonstrate numerically that moderate level of anticipation generically promotes convergence towards consensus, while too high level disturbs it. Motivated by this observation, we derive sufficient conditions for asymptotic consensus in the multiple-agent systems, which are explicit in the parameter values delay length and anticipation level, and independent of the number of agents. The proofs are based on construction of suitable Lyapunov-type functionals.