Cooperation in a Noisy Room - Multi Agent Reinforcement Learning in the Presence of Communication Noise
Salim Dridi · 2024
Collective intelligence is a crucial feature of a variety of living species. It has enabled them to survive in a wide variety ofenvironments around the world. Collective intelligence is necessary to handle difficult survival concerns, from simpleorganizations in ant colonies to sophisticated systems in human society. Humans, too, are social animals who share theirexperiences with one another, and this communication helps them evolve in such a way that they do not make the samemistakes as other humans.When it comes to intelligent agents, the same scenario is repeated. Intelligent agents can operate in both independentand cooperative modes. The goal of this study is to compare the two systems and see how the efficiency of a systemchanges when cooperative agents instead of autonomous ones are used. After that, we’ll look at how these cooperativeagents communicate in an ideal environment and in an environment where noise is present in their communication.This study will look at how agents behave in noisy situations and how their communication breaks down whenthe signal-to-noise ratio rises. This study describes how cooperative agents outperform independent agents, what analogies amulti-agent system might use to interact, how such systems connect to a human-like scenario, and how all of these elementscan be implemented.