Assessment of Estimation Methods for Demolition Waste Volume and Cost

For the past few decades, researchers have tried to make a sustainable built environment by maximizing recycling and reuse of construction and demolition (C&D) waste. In particular, demolition waste accounts for more than 90% of the total C&D waste generated in the US, thus signifying substantial potential for recycling and reuse. While there have been several models to estimate construction waste available for supporting its waste reduction planning, however, there has been a lack of estimation models for demolition waste. This research seeks to evaluate and compare the feasibility and accuracy of four common estimation approaches for demolition waste (i.e., demolition waste volume and cost): a linear regression, an artificial neural network, and two advanced case-based reasoning approaches, which utilize several regression models on selected instances to improve the overall accuracy of predicted cost and volume of demolition waste. A database of 52 demolition projects, containing information on architectural characteristics, permit history, value, and contract requirements, is used to train models and facilitate evaluation. Different estimation methods are compared in terms of estimation accuracies while discussing the potential improvement of each method. This study will serve as the cornerstone to develop a more reliable demolition estimation model in the future.