Evolutionary and Coevolutionary Multi-Agent Design Choices and Dynamics

2025-07-07

AI Summary

This study investigates multi-agent adversarial scenarios in cybersecurity, focusing on the interplay between controller representations—specifically, code-logic syntax versus alternative representations—and evolutionary algorithms. We propose a large language model (LLM)-enhanced mutation operator and systematically compare unilateral evolution (optimizing one agent only) against coevolution (joint optimization of both agents) within a syntax-guided evolutionary framework. Experimental results demonstrate that code-logic syntax representation combined with LLM-augmented evolution achieves optimal team-level performance. Unilateral evolution yields higher and more stable performance peaks, whereas coevolution suppresses extreme performance but introduces training instability and dynamic fluctuations. The core contribution lies in uncovering the coupled influence of representation design, evolutionary paradigm, and LLM-assisted mechanisms on multi-agent collaborative learning efficacy.

Technology Category

• Multiagent Systems: Adversarial Agents
• Machine Learning: Evolutionary Learning
• Game Theory and Economic Paradigms: Adversarial Learning

Application Category

• Web Mining and Content Analysis: Models for Web evolution
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• User Modeling, Personalization and Recommendation: Attacks and countermeasures in recommendation systems

Abstract

We investigate two representation alternatives for the controllers of teams of cyber agents. We combine these controller representations with different evolutionary algorithms, one of which introduces a novel LLM-supported mutation operator. Using a cyber security scenario, we evaluate agent learning when one side is trained to compete against a side that does not evolve and when two sides coevolve with each other. This allows us to quantify the relative merits and tradeoffs of representation and algorithm combinations in terms of team performance. Our versions of grammatical evolution algorithms using grammars that allow a controller to be expressed in code-like logic can achieve the best team performance. The scenario also allows us to compare the performance impact and dynamics of coevolution versus evolution under different combinations. Across the algorithms and representations, we observe that coevolution reduces the performance highs and lows of both sides while it induces fluctuations on both sides. In contrast, when only one-side is optimized, performance peaks are higher and is more sustained than when both sides are optimized with coevolution.

Problem

• Compare controller representations for cyber agent teams
• Evaluate coevolution vs single-side evolution in cyber security
• Assess performance of LLM-supported mutation in evolutionary algorithms

Innovation

• LLM-supported mutation operator in evolution
• Grammatical evolution with code-like logic
• Coevolution dynamics reduce performance extremes