Conceptual cost-prediction model for public road planning via rough set theory and case-based reasoning

Long-term transportation policies require government officials to predict the cost of public road construction during the conceptual planning phase. However, early cost prediction is often inaccurate because public officials are not familiar with cost engineering practices, and moreover, have limited time and insufficient information for estimating the possible range of the cost distribution. This study develops a conceptual cost prediction model by combining rough set theory, case-based reasoning, and genetic algorithms to better predict costs in the conceptual planning phase. Rough set theory and qualitative in-depth interviews are integrated to select the proper input attributes for the cost prediction model. Case-based reasoning is then applied to predict road construction costs by considering users' difficulties in the conceptual policy planning phase. A genetic algorithm is also used to assist the rough set model and case-based reasoning model to obtain optimal solutions. The result of the analysis shows that the proposed conceptual cost prediction model is reliable and robust compared to the existing cost prediction model.