Intelligent Decision Support Systems Leveraging Knowledge Graphs and Ontologies for AI-Empowered Intelligent Engineering Systems

Abstract

Integrating knowledge graph and ontology into intelligent decision support system is a great progress in making decisions in complex engineering environment. With the engineering system theory, network physical system principle and information theory, these semantic technologies provide organized knowledge that machines can understand. This helps to make decisions that take into account the context, can be explained, and make full use of resources. This paper focuses on the theoretical basis, how to mix methods, where to use them, and the problems of ontology and knowledge map in intelligent decision-making system and intelligent engineering system. The technologies that help to achieve this goal include: digital twins that link the virtual with the reality, reinforcement learning that uses semantic rules to organize tasks, computer vision with ontology annotation to find defects in real time, and edge computing to conduct rapid semantic reasoning. Specific examples of intelligent manufacturing, energy system and infrastructure monitoring show that measurable progress has been made in work efficiency, reducing uncertainty and reusing knowledge. Critical analysis compares these benefits with limitations, such as the work of maintaining ontology, the difficulty of large-scale reasoning, the problem of data quality, and the vulnerability when concepts change. An initial idea put forward an adaptive framework called semantic network physical system, in which there is a meta-semantic layer to measure the usefulness of knowledge and make the system develop itself. The conclusion is that different disciplines need to work together to create a reliable and highly autonomous engineering system.