Neural network modeling for rework related cost overrun and contractual claims in construction projects

The dynamics of rework related transactions in construction projects become very complicated because of various influencing factors such as multiple stakeholder interactions and overlapping interfaces. In order to understand the significance of rework based impacts on different performance related aspects in construction projects (e.g. cost overrun, time overrun, contractual claims), a pilot study was recently launched in Hong Kong. The knowledge-mining exercise aimed to consolidate the rework experiences from various recently completed construction projects, and this mainly included (i) a set of exploratory interviews and (ii) a questionnaire survey. It was considered that artificial neural network modeling approaches can be developed for mapping rework related impacts on different aspects of project performance. Applications of advanced neural network architectures such as General Regression Neural Networks (GRNN) have been explored for modeling rework based cost overrun and contractual claims in construction projects. A consolidated summary of initial findings from the neural network modeling for rework related cost overrun and contractual claims is presented in this paper.