Enhancing project completion date prediction using a hybrid model: rule-based algorithm and machine learning algorithm

The necessity for this research arises due to the pressing need to create reliable early indicators to accurately project the completion dates, i.e., First Gas Date (FGD) and First Oil Date (FOD), of Malaysia's upstream oil and gas industry. The central purpose of this research is to significantly increase the predictability of these milestone dates, thereby eliminating the risks associated with high and dynamic fluctuations in schedules. The study employs a hybrid predictive model that combines Big Data technologies, Extract Load Transfer (ELT) processes, rule-based algorithms (RBA), machine learning (ML), and Power BI visualizations. The methodologies employ strict checks based on the 14-point Defense Contract Management Agency (DCMA) schedule health framework to detect issues in the schedules at an early stage and maintain the integrity of the schedules. The results demonstrate that the integration of Big Data analytics, ELT pipelines, and machine learning models effectively translates complex sets of Primavera P6 EPPM into actionable insights, providing accurate dynamic forecasts and warning of delays in advance. The research finds that the implementation of the hybrid framework ensures the on-time completion of project milestones, stabilizing scheduled FGD and FOD dates. The findings of this study have significant implications for National Oil Companies (NOCs) and operators, facilitating proactive intervention strategies, enhancing transparency in decision-making, and supporting the early monetization of resources. Ultimately, this method improves the protection of project schedules, optimizes infrastructure utilization, and maximizes economic returns and strategic alignment in Malaysia's energy industry.