A Three-Level Agent-Based Framework For Resource Allocation Optimization In Flexible Manufacturing Systems

Abstract

Agent-based approach has become a prominent paradigm for analysis, design and implementation of modern industrial systems. The paper focuses on the development of practical applications to optimize resource allocation in flexible manufacturing systems using auto-guided vehicles (AGVs). The system uses Agent Unified Model Language (AUML) to design models for various agents such as machines, vehicles, and products. In the first level, we define the role, capacity, and function of the agent in the system and describe its functions in detail. The second level deals with the behavior of these agents and explains in detail their decision-making processes and tasks. The third level examines how agents interact and communicate to achieve system-wide objectives. Using Contract Net Protocol (CNP), agents can demonstrate and manage interactions, ensuring efficient negotiations and resource allocation. The multi-level approach offers effective solutions for optimizing operations of dynamic and distributed manufacturing systems.