Heuristic Driven Hybrid Analytical-Artificial Intelligence Concept for Risk Prediction in Construction Megaprojects

Abstract

In this paper, a cross analytical-Artificial Intelligence (AI) solution for target centric risk identification and clustering has been developed for better project planning and risk management. A mixed research paradigm has been initially applied in order to identify the optimal set of risk factors based on the opinion gathered from experienced subject matter experts. The statistical analysis has revealed the key factors adversely affecting project objectives. In order to enable target centric risk prediction, the processed data have been exploited by the proposed Heuristic Driven Clustering-based Target-Centric Risk Prediction Model for Construction Megaprojects (HCTCRP). First, the proposed model has synthesized a set of 40000 samples containing the different risk factors and their variation with respect to the project endeavors. Subsequently, by processing significant predictor test and Firefly algorithm based Fuzzy C-Means clustering (FAFCM), the proposed model has predicted the high-risk factors and the sub-risk components. This significantly contributes to the knowledge base and may be useful for better risk management.