Estimating construction waste generation in residential buildings: A fuzzy set theory approach in the Brazilian Amazon

The estimate of construction waste generation is the key decision-making information for policy-makers, construction managers, and the like to devise informed waste management strategies. However, estimating construction waste generated from projects is particularly onerous, as numerous factors related to design, site, and construction are largely in a fuzzy nature when the estimating job is conducted. Built upon previous studies, this paper seeks to develop a model that can be used to estimate construction waste generation based on fuzzy set theory. It follows a trilogy of methodology, including model development, sensitivity analysis, and model validation. A set of IF-THEN rules are developed based on two independent variables, built area and number of floors. A sensitive analysis was conducted to evaluate the influence of the independent variables on waste generation. The model is further calibrated and verified through a case study of 23 residential buildings constructed in the Brazilian Amazon. The model obtained an accuracy of 64.29% in the development phase and 66.67% in the validation phase, showing that the results are largely acceptable. By using this model, it is possible for a waste manager to draw up a baseline graph to indicate the volume of construction waste generation as his/her building project as it progresses. The research is also of novelty by using fuzzy set theory to deal with the fuzzy nature of waste generation in construction projects. Further studies are recommended to enhance the accuracy level of the model by engaging more factors and more quality data. •The paper aims to develop a model to quantify waste generation in the Amazon, Brazil.•The model considers the fuzziness of factors that could impact waste generation.•Number of floors, and construction floor areas are critical factors.•The model obtained an accuracy of 64.29% in the development phase.•The model accuracy can be improved by engaging more factors and more quality data.