English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

In this work we consider a generalization of the well-known multivehiclerouting problem: given a network, a set of agents occupying a subset of itsnodes, and a set of tasks, we seek a minimum cost sequence of movements subjectto the constraint that each task is visited by some agent at least once. Theclassical version of this problem assumes a central computational server thatobserves the entire state of the system perfectly and directs individual agentsaccording to a centralized control scheme. In contrast, we assume that there isno centralized server and that each agent is an individual processor with no apriori knowledge of the underlying network (including task and agentlocations). Moreover, our agents possess strictly local communication andsensing capabilities (restricted to a fixed radius around their respectivelocations), aligning more closely with several real-world multiagentapplications. These restrictions introduce many challenges that are overcomethrough local information sharing and direct coordination between agents. Wepresent a fully distributed, online, and scalable reinforcement learningalgorithm for this problem whereby agents self-organize into local clusters andindependently apply a multiagent rollout scheme locally to each cluster. Wedemonstrate empirically via extensive simulations that there exists a criticalsensing radius beyond which the distributed rollout algorithm begins to improveover a greedy base policy. This critical sensing radius grows proportionally tothe $\log^*$ function of the size of the network, and is, therefore, a smallconstant for any relevant network. Our decentralized reinforcement learningalgorithm achieves approximately a factor of two cost improvement over the basepolicy for a range of radii bounded from below and above by two and three timesthe critical sensing radius, respectively.

Distributed Online Rollout for Multivehicle Routing in Unmapped Environments