$\mathrm{TRIZ}^{a}$: Guiding Agent Evolution from Pattern Recognition to Solution Invention

2026-10-03

AI Summary

This study addresses the lack of theoretical guidance and interpretability inherent in random mutation within R&D automation by proposing an evolutionary search framework that integrates TRIZ theory with large language model agents. Specifically, this method replaces stochastic mutations with TRIZ inventive principles, leverages the contradiction matrix to guide solution generation, and incorporates functional information alongside ideality-based quantitative evaluation and a hard-gating verification mechanism. Evaluated across multiple cybersecurity benchmarks, the proposed approach significantly improves F1 scores while ensuring full-process traceability and rigorous validation. Ultimately, this work establishes a novel paradigm for structured and interpretable R&D automation.

Abstract

We propose $\mathrm{TRIZ}^{a}$ (TRIZ exponentiated by an agent), a general R\&D automation paradigm that combines TRIZ inventive theory with LLM-driven agent evolutionary search. TRIZ's 40 inventive principles and contradiction matrix provide structured, explainable directions for solution generation, replacing random or untyped mutation with theory-guided ideation. Functional information (FI), operationalized under a frozen reference contract, is combined with TRIZ Ideality to measure useful and harmful function on a commensurable information scale, while hard gates keep promotion distinct from metric improvement. We validate $\mathrm{TRIZ}^{a}$ in cybersecurity--an adversarial and rapidly evolving domain--on PowerDuck GOOSE, CICIoT2023, and CIC-DDoS2019. Under paired-rerun protocols with protocol fingerprinting and hard-gate validation, the legacy experiments yield absolute F1 improvements of $+2.88$, $+4.23$, and $+0.15$ percentage points, respectively. A completed 45-activity CICIoT2023 campaign further increases macro-F1 from $0.8325$ to $0.8483$, but does not pass its frozen promotion gate. Every result remains traceable from contradiction identification and TRIZ principle selection to code transformation, evaluation metrics, and promotion decision.

Problem

• Agent Evolution
• R&D Automation
• TRIZ
• Cybersecurity
• Large Language Models

Innovation

• TRIZ
• LLM-driven Agent
• Evolutionary Search
• Functional Information
• R&D Automation