Complexity identification in major infrastructure project information systems using graph theory

Abstract

AbstractInformation management for major infrastructure projects is a critical task requiring ever increasing time and resources. Whilst the digitisation of construction industry information has progressed over multiple dimensions, there are still gaps in machine readable information for adequate suitability assessment of projects against their requirements. With digital twins at the heart of the digitisation agenda, it is vital that the industry expand the information available to validate and verify asset and network outcomes throughout their lifecycle. Within the construction industry, systems engineering is being deployed for asset and information requirements and model‐based systems engineering can support digital techniques in managing construction information requirements. However, as an intensive task it must be selectively deployed to obtain value within a complex operating system. An opportunity exists to use data and relationships held in electronic document management systems that currently store project requirements. To achieve this, graph/network theory can be used to visualise and analyse the connectivity of document datapoints. This study deploys a python modelling environment to create digraphs that are used to visualise and identify key data from a document management system, identifying the highest degree and betweenness for documents from a sample size of 20 queries producing 327 vertices. It demonstrates a useful means of interrogating the data rapidly that is quick to scale and expand. There are opportunities to deploy this across multiple digital systems to further map complexity and interactions between them. Identification is the critical first step for data driven point of application ‘surgical’ model‐based systems engineering.