RebecAI: An Intelligent Workflow for Formal Verification and Model Checking of Rebeca Based on the Actor Model

The RebecAI project combines AI agents with the Rebeca Actor model, aiming to automate formal verification and model checking of distributed systems. By integrating AI capabilities, the project lowers the barrier to formal verification, improves verification efficiency, and provides a more reliable quality assurance method for distributed system development.

Overview of the Actor Model

The Actor model is a classic framework for concurrent computing. As the basic unit, an Actor has the characteristics of encapsulating state, message passing, dynamic creation, and behavior switching.

Rebeca Language

Rebeca is a modeling language based on the Actor model, providing formal syntax and semantics to support modeling and verification of concurrent systems.

Value of Formal Verification

• Exhaustive Guarantee: Traverse all states to find boundary cases missed by testing;
• Early Defect Detection: Identify issues like race conditions and deadlocks during the design phase;
• Specification Precision: Eliminate ambiguities in natural language and accurately describe system behavior.

Intelligent Modeling Assistance

AI assists in requirement understanding (converting natural language to formal specifications), model generation, code completion, and pattern recognition.

Automated Verification Workflow

1. Property Extraction: Extract verification properties from requirement documents;
2. Specification Conversion: Convert to LTL/CTL temporal logic formulas;
3. Detection Strategy: Intelligently select algorithms and parameters;
4. Result Analysis: Interpret results and generate repair suggestions.

Counterexample Explanation and Repair

AI explains counterexample execution sequences, analyzes root causes, provides repair solutions, and assists in verifying the repaired model.

Skill Modularization

Decompose the verification workflow into independent skill units such as model parsing, state space analysis, property template management, and report generation.

Toolchain Integration

Compatible with Rebeca model checkers (e.g., Modere), supports multiple verification backends, and has an extensible plugin architecture.

Interactive Experience

Provides an intuitive interface for incremental development, real-time feedback, hypothesis exploration, and collaborative review.

RebecAI is applicable to various distributed system verification scenarios:

• Communication Protocol Verification: Ensure reliability and order of message transmission (e.g., distributed consensus protocols);
• Embedded System Analysis: Verify real-time constraints and resource competition;
• Cloud Service Reliability: Model microservice architectures and verify mechanisms like fault recovery;
• Smart Contract Auditing: Detect vulnerabilities in blockchain contracts such as reentrancy attacks and race conditions.

Current Challenges

• State Space Explosion: Complex systems have huge state spaces, requiring a combination of abstraction and AI-guided strategies;
• Formalization Barrier: Reducing the learning curve requires better educational tools and interactive design;
• Scalability: Supporting industrial-grade systems requires distributed detection and cloud computing integration.

Future Directions

Integrate multi-modal inputs such as code and architecture diagrams to provide more comprehensive intelligent assistance.

RebecAI represents a cutting-edge exploration of the integration of AI and formal methods. By introducing large language model capabilities, it significantly lowers the barrier to formal verification and improves efficiency. With the development of the project, more AI-driven verification models are expected to emerge, promoting the quality improvement of distributed system development.