Adaptive predictive neural network for brickwall production rates
Abstract
The productivity of brickwork is an important factor in the construction industry, as it directly
affects project timelines and costs. In this study, we investigated the relationship between the
production rate of brickwork and various factors such as wall height, length, crew configuration
and crew rates and area, wall length, height, area built, and the number of masons were the most
influential factors. The data was collected from 30 different construction sites, and the production
rate was measured in terms of square meters of wall constructed and also the square area of walls
plastered per day. We used a Weibull distribution to model the data, as it is well-suited for handling
censored data. The results showed a strong positive linear relationship between the production rate
and wall height, length, and area, indicating that an increase in these factors leads to an increase
in productivity. Furthermore, we developed an artificial neural network (ANN) model to predict
the production rate of brickwork based on these factors, the R-squared value for training,
validation, testing, and overall data instance mapping was 86.8%, 54.7%, 79.5%, and 84.9%
respectively. The model architecture was composed of a neuron-hidden layer combination of 4
hidden layers and 9 neurons with an MSE of 0.053 signifying a 95% model accuracy The findings
of this study can be used to improve the productivity of brickwork in construction projects and
inform decision-making processes.