Using Intelligent Case-Based Reasoning (CBR) for Parametric Cost Estimating in Construction

Estimating project cost can be fairly challenging during the project's early stages due to the limited job information and details. This type of estimates, which are typically called parametric cost estimates, embarks on historical records and expert judgment to perform the costing function. Over the past two decades, attempts have been made to use artificial intelligence (AI) technologies, such as neural networks, for such purpose with varying levels of success. The paper investigates the suitability of another contemporary and innovative AI technique known as Case-based Reasoning (CBR) for the same. CBR can search an existing database/library of past projects for the most relevant cases and perform the costing of a new job based on the most relevant project cases. Study exemplifies this innovative approach in highway construction projects in Egypt. For implementation, 23 drivers of project cost have been identified, based on literature review and unstructured interviews with subject matter experts in Egypt. Data pertinent to these cost drivers for 36 highway projects was collected and used for constructing the CBR case library. The retrieval and adaptation approaches are employed to extract useful knowledge embedded in the case library and then use for estimating the cost of a new highway project. Automation of the retrieval and adaptation process was done through the software CBR Works TM version 4.0. CBR shows promise in estimating the costs of tested cases. Paper concludes with comments on the parametric cost estimating function and recommendations on the usage of innovative approaches such as CBR in performing it