Dynamic max-consensus with local self-tuning

This work describes a novel control protocol for multi-agent systems to solve the dynamic max-consensus problem. In this problem, each agent has access to an external timevarying scalar signal and has the objective to estimate and track the maximum among all these signals by exploiting only local communications. The main strength of the proposed protocol is that it is able to self-tune its internal parameters in order to achieve an arbitrary small steady-state error without significantly affecting the convergence time. We employ the proposed protocol in the context of distributed graph parameter estimations, such as size, diameter, and radius, and provide simulations in the scenario of open multi-agent systems. Copyright (C) 2022 The Authors.