Autonomous Decision Systems for Ubiquitous Smart Factories

Abstract

Intelligent manufacturing is a fundamental aspect of the evolution of the smart factory. Automated decision-making systems leverage intelligence to simplify, manage, and operate manufacturing operations flexibly. At this stage, these systems, when combined, incorporate real-time processing, big-data predictive analytics, and cyber-physical systems to enable real-time decision-making and seamless service delivery. Smart factories would seemingly operate independently and provide near-real-time, responsive operations to changes in production schedules, opportunity and risk levels for resource allocation and repurposing, maintenance shutdowns, and the optimization of morning and afternoon idle time, etc. Despite the practical value of both predictive analytics and decision-making systems, more deliberate attention is being paid to the sociotechnical vulnerabilities, the integration of legacy systems with decision-making, and the ethics of automated decision-making. Scoping reviews clearly refer to "automated decision-making systems" primarily as a smart factory advancement, with careful evaluation of the technologies that support them. Case studies in the literature demonstrate the value of operational autonomous decision-making systems for manufacturing industries. Here, literature gaps are addressed, and clear next steps are outlined to drive system flexibility, automation ethics, and interoperability and cross-system resilient frameworks. This is important because having autonomous decision systems is one of the most relevant steps to achieving Industry 4.0 and more advanced stages. These systems will enable the establishment of responsive, sustainable, and adaptive ecosystems in smart factories.