New Results on Finite Convergence Time Mode Consensus: a Summary

This paper studies the distributed mode consensus problem in a multi-agent system. Three algorithms are proposed to find the most frequent attribute (the mode) owned by the agents via distributed computation. The first algorithm computes the frequency of each attribute using consensus protocols rooted in blended dynamics, then identifies the most frequent attribute as the mode. The second algorithm, under the assumption that each agent possesses a priori knowledge of a minimum frequency for the mode, can decrease the frequency computations required at each agent for large lower bounds. In contrast, the third algorithm eliminates the necessity for such information by implementing an adaptive updating mechanism. These algorithms successfully determine the mode within a finite time frame, and predictive estimates for convergence time are included. Moreover, the first and second algorithms demonstrate plug-and-play property with a dwell time.