Designing a Multi-Agent PLM System for Threaded Connections Using the Principle of Isomorphism of Regularities of Complex Systems

Abstract

The PLM concept implies the use of heterogeneous information resources at different stages of the product life cycle, the joint work of which allows the user to effectively solve the problems of product quality and various costs. According to the principle of isomorphism of complex systems’ regularities, an effective PLM system must have these regularities. Unfortunately, this principle is not often fundamental when designing PLM systems. The purpose of this paper is to develop principles, based on the general theory of systems, for the design, operation and use of PLM systems and the implementation on their basis of the educational framework of a PLM system for threaded connections with the possibility of its effective development, research and study. The multi-agent approach to the development of a PLM system provides the necessary prerequisites for the emergence of system-wide regularities in it. The parallel work of agents is implemented using the actor model and the Ray Python-package. Agents for the logical inference of knowledge base facts, CAD/FEA/CAM/SCADA agents, agents for optimization by various methods, and other agents have been developed. Open-source software was used to develop the system. Each agent has relatively simple behavior, implemented by its rule function, and can interact with other agents. The system can work in interactive mode with the user or in automatic mode according to a simple algorithm: the rule functions of all agents are executed until at least one of them returns a value other than None. Examples of the system operation are given and system-wide regularities such as emergence, historicity and self-organization are demonstrated in it.