MermaidFlow: Redefining Agentic Workflow Generation via Safety-Constrained Evolutionary Programming

2025-05-29

AI Summary

Existing LLM-driven workflow generation methods often produce fragile, non-executable plans. To address this, we propose a Mermaid-based, verifiable intermediate representation and establish a graph-structured evolutionary paradigm under safety constraints: domain-aware evolutionary operators—including crossover, mutation, insertion, and deletion—jointly optimize semantic correctness and structural diversity within a statically verifiable workflow search space. Our method requires no fine-tuning and supports zero-shot workflow optimization. On agent reasoning benchmarks, it significantly improves workflow executability (+23.6%) and convergence speed (1.8× acceleration), delivering consistent performance gains without modifying task configurations. The core innovation lies in the first integration of safety constraints directly into the graph-structured evolutionary process, thereby unifying verifiability, robustness, and generalization.

Technology Category

• Search and Optimization: Metareasoning and Metaheuristics
• Planning, Routing, and Scheduling: Planning with Language Models
• Humans and AI: Human-Aware Planning and Behavior Prediction

Application Category

• Graph Algorithms and Modeling for the Web: Foundation models and LLMs for Web-related graphs
• Semantics and Knowledge: Methods to enhance, augment, integrate or synergize semantic models such as knowledge graphs and LLMs
• Economics, Online Markets and Human Computation: Architectures and workflows that use LLMs for crowd work

Abstract

Despite the promise of autonomous agentic reasoning, existing workflow generation methods frequently produce fragile, unexecutable plans due to unconstrained LLM-driven construction. We introduce MermaidFlow, a framework that redefines the agentic search space through safety-constrained graph evolution. At its core, MermaidFlow represent workflows as a verifiable intermediate representation using Mermaid, a structured and human-interpretable graph language. We formulate domain-aware evolutionary operators, i.e., crossover, mutation, insertion, and deletion, to preserve semantic correctness while promoting structural diversity, enabling efficient exploration of a high-quality, statically verifiable workflow space. Without modifying task settings or evaluation protocols, MermaidFlow achieves consistent improvements in success rates and faster convergence to executable plans on the agent reasoning benchmark. The experimental results demonstrate that safety-constrained graph evolution offers a scalable, modular foundation for robust and interpretable agentic reasoning systems.

Problem

• Fragile unexecutable plans from unconstrained LLM-driven workflow generation
• Lack of safety-constrained methods for agentic workflow evolution
• Need for verifiable intermediate representation in agentic reasoning systems

Innovation

• Safety-constrained graph evolution for workflows
• Mermaid-based verifiable intermediate representation
• Domain-aware evolutionary operators for correctness