2019/11/06 by Shirantha Welikala, Welikala, Shirantha, Christos G. Cassandras +1
Computer Science · Earth and Planetary Sciences · #Age of Information Optimization #Distributed Control Multi-Agent Systems #Energy Efficient Wireless Sensor Networks #FOS: Computer and information sciences #FOS: Mathematics #Marine and coastal ecosystems #Multiagent Systems (cs.MA) #Optimization and Control (math.OC)
paper · pdf · doi:10.48550/arxiv.1911.02658
openalex publication_date 2019/11/06 · openalex created_date 2022/07/26 · openalex updated_date 2026/07/28
This paper considers the optimal multi-agent persistent monitoring problem\ndefined for a team of agents on a set of nodes (targets) interconnected\naccording to a fixed network topology. The aim is to control this team so as to\nminimize a measure of overall node state uncertainty evaluated over a finite\ntime interval. A class of distributed threshold-based parametric controllers\nhas been proposed in prior work to control agent dwell times at nodes and\nnext-node destinations by enforcing thresholds on the respective node states.\nUnder such a Threshold Control Policy (TCP), an on-line gradient technique was\nused to determine optimal threshold values. However, due to the non-convexity\nof the problem, this approach often leads to a poor local optima highly\ndependent on the initial thresholds used. To overcome this initialization\nchallenge, we develop a computationally efficient off-line greedy technique\nbased on the asymptotic analysis of the network system. This analysis is then\nused to generate a high-performing set of initial thresholds. Extensive\nnumerical results show that such initial thresholds are almost immediately\n(locally) optimal or quickly lead to optimal values. In all cases, they perform\nsignificantly better than the locally optimal solutions known to date.\n