Workshop Proceedings Oil platforms, classified as critical infrastructure, require intricate regulations and standards to ensure their safety. Constructing oil platforms involves various stakeholders, such as major oil companies like Equinor, Engineering, Procurement, and Construction (EPC) service providers like Aibel, and local governments. Consequently, a significant volume of documents is generated and exchanged among these parties to facilitate data and information transfer. However, ensuring data quality and managing operational costs, including time and human resources, prove challenging in project execution. To tackle these issues, Aibel has developed the Material Master Data (MMD) ontology, which serves as a standardized vocabulary for product classification and project description.
CEUR
ceur-ws.org Introduction CEUR Workshop Proceedings brought substantial benefits, substantially decreasing project costs by reducing the procurement of incorrect pipe components. By leveraging the MMD ontologies, we could enhance the ability to ensure compliance, minimize errors, and optimize cost-eficiency in the pipe component ordering process. This development highlights the positive impact that reasoning on large ontologies can have on industrial operations and financial outcomes.
Solution In the aforementioned use cases, only a small portion of the ontology is typically involved. For instance, when modifying ontologies or querying the entire ontology, it is unnecessary to perform reasoning tasks on the entire ontology. Instead, a subset of the ontology can be extracted and reasoning can be conducted on this smaller set, yielding the same reasoning results. By reducing the size of the ontology, the reasoning tasks naturally become faster compared to reasoning on the entire ontology. This technique is called ontology modularity. An ontology module refers to a subset of the ontology that preserves the complete logical consequences relative to the given vocabulary, called a signature. Various module notions have been proposed, including locality-based modules, semantic modules, deductive modules, and more, tailored for diferent use cases. In our paper, we use locality-based modules due to their speed in module computation and ease of use. Although they may contain redundant axioms, these modules retain all the logical consequences of the original ontology. By leveraging ontology modularity, we can enhance reasoning eficiency by working with smaller subsets of ontology while preserving the necessary logical information.
Our framework comprises three key steps to enhance ontology reasoning eficiency: 1. Signature extraction: We identify a set of classes and properties crucial for ontology modelling and queries, which allows us to focus on the essential elements of the ontology. 2. Module extraction: Using the extracted signature, we generate a subset of the ontology called the locality-based module. This module includes all relevant axioms and logical consequences related to the specified signature. By reducing the ontology’s size, we retain the necessary information for reasoning while significantly improving eficiency. 3. Reasoning on the modules: We perform reasoning tasks exclusively on the extracted module. This approach enables faster and more eficient reasoning compared to working with the entire ontology, while still producing the same high-quality results.