Consensus among Learning Agents: A Multi-AgentReinforcement Learning Framework withGame-Theoretic Incentives

Abstract excerpt

<title>Abstract</title> <p>Consensus protocols for blockchain systems are typically designed around a fixedparticipant model: rational miners, rational stakers, or protocol-scheduled valida-tors whose strategy space is bounded by rules set offline. As autonomous agentsincreasingly act on behalf of humans in decentralized settings, the participantsthemselves become learning entities whose policies drift during ope...

Topics

Open a Topic to create a Post that cites this publication.

• Ethics in Clinical Research
• Evolutionary Game Theory and Cooperation
• Evolutionary Psychology and Human Behavior
• Genetic Variation
• Mathematical and Theoretical Epidemiology and Ecology Models
• Neuroethics, Human Enhancement, Biomedical Innovations
• Racial Groups

Identifiers and source

Literature Corpus work

65c14eb4-de30-5196-b5a0-5bf1b8dc9b4a

DOI

10.21203/rs.3.rs-10013916/v1