A Comparative Study of Human-Inspired Communication Strategies in Multi-Agent Environments

The effective coordination of autonomous agents in shared, resource-limited environments requires robust communication strategies. Although various methods exist for multi-agent systems (MAS), strategies inspired by human communication remain largely unexplored. This research explores the impact of human-inspired communication strategies on agent performance within multi-agent systems (MAS), using the NetLogo Sugarscape 2 model. The study compares Proximity, Small Group Communication, and Deception with a baseline of no communication and a Q-learning method, where agents adapt their communication. The aim was to evaluate if human-like strategies improve agent performance in resource-scarce environments. Results showed that Proximity and Small Group Communication led to higher survival rates and resource acquisition, while Deception hindered performance, indicating that dishonesty can impede resource competition. Although Q-learning demonstrated adaptability, its effectiveness decreased in larger populations, where simpler methods were more successful. Overall, the study highlights that human-inspired strategies can influence agent interactions, but may not exceed existing MAS methods. This work emphasizes the importance of balancing simplicity and adaptability when designing effective communication protocols for artificial agent societies, and it contributes to the field of complex social systems modelling.

• Publication year

  2025

• Venue

  Proceedings of the 2025 International Conference on Artificial Intelligence, Big Data, Computing and Data Communication Systems

• Publication date

  2025-11-25

• Fields of study

  Not labeled

• Identifiers
  DOI 10.1145/3759023.3759094
• External record

  Open on Semantic Scholar

• Source metadata

  Semantic Scholar

• No claims are published for this paper.

• No concepts are published for this paper.

• A Survey of Multi-Agent Reinforcement Learning with Communication

  2022cited by this paper
• Computational Experiments for Complex Social Systems—Part I: The Customization of Computational Model

  2022cited by this paper
• Benchmarking Multi-Agent Deep Reinforcement Learning Algorithms in Cooperative Tasks

  2020cited by this paper
• A survey and critique of multiagent deep reinforcement learning

  2018cited by this paper
• TarMAC: Targeted Multi-Agent Communication

  2018cited by this paper
• Agent-Based and Individual-Based Modeling: A Practical Introduction

  2011cited by this paper
• CO-EVOLUTION OF NEOCORTEX SIZE , GROUP SIZE AND LANGUAGE IN HUMANS

  2008cited by this paper
• Communication Scheme Comparison for a Distributed Multi-Agent System

  2008cited by this paper
• Social semantics: how useful has group selection been?

  2008cited by this paper
• Social semantics: altruism, cooperation, mutualism, strong reciprocity and group selection

  2007cited by this paper
• Evolution of indirect reciprocity by image scoring

  1998cited by this paper
• Social Brain Hypothesis

  1998cited by this paper
• Reinforcement Learning: An Introduction

  1998cited by this paper
• Growing Artificial Societies: Social Science from the Bottom Up

  1996cited by this paper
• Q-learning

  1992cited by this paper
• The Strength of Weak Ties

  1973cited by this paper
• The Tragedy of the Commons

  1968cited by this paper

• No citing papers are available for this paper.