Multiagent System for Discrete Manufacturing with Integration into the MES/SCADA System and Production Decision Support

Abstract

Discrete manufacturing faces significant challenges related to high variability, mass customization, and the demand for rapid responses to operational failures, which increase costs and compromise process efficiency. Industry 4.0 emerges as a response to this scenario by promoting digitalization and intelligent system integration, enabling real-time monitoring and dynamic adaptation of production. The integration between the Manufacturing Execution System (MES) and Multi-Agent Systems (MAS) becomes essential to optimize processes, enable greater decision-making autonomy, and increase the resilience of the production system. This work proposes a multi-agent approach to optimize production in discrete manufacturing environments, integrating with MES/SCADA systems and establishing the foundations for a Digital Twin with Artificial Intelligence (AI) capabilities. The methodology was applied at the Advanced Manufacturing Plant (PMA) of SENAI CIMATEC, a demonstrative Industry 4.0 environment. The initial steps involve defining the agents, modeling the system, and integrating with MES, using a Service-Oriented Architecture (SOA) to expose functionalities as microservices. The results include the mapping and classification of PMA components, as well as the proposal of a preliminary agent structure, modeled in UML and divided into two groups: Device Agents, responsible for field data collection, and Management and Analysis Agents. It is expected that the Analyzer Agent will play a key role in detecting deviations or potential failures in the process by employing AI to predict faults and suggest rescheduling. The proposed solution offers a conceptual basis that integrates traditional technologies and intelligent agents to build an AI-enabled Digital Twin, capable of supporting dynamic rescheduling and adaptive control, enhancing responsiveness to changes and unforeseen events.