Construction industry is a labor-based industry. The productivity of construction workers is complex and can have significant effects on the construction industry, government, and the overall economy. To maintain profitability, the construction industry must always aim for increased productivity through the identification and management of the factors that influence it, as well as routine checks to see if worker productivity within their company is meeting criteria. The primary aim of this research is to investigate labor productivity of brickwork within the building construction sector in Kathmandu Metropolitan City. Additionally, it aims to predict the labor productivity of brickwork in the context of building construction within Kathmandu Metropolitan City using artificial neural networks (ANN) method. Out of sixty questionnaires distributed, fifty-four respondents, twenty-eight from contractors, seventeen from consultants and nine from client responded giving their views for the ranking of factors affecting labor productivity. Listing of the influencing factors and pretesting of the questionnaire was done prior to the distribution of the questionnaire in the selected projects. Relative importance index technique was used to calculate the rank of the factors. Eighteen most impacting factors from the fifty-five listed factors were chosen and encoding of these factors were done to develop the artificial neural networks (ANN) in SPSS Statistics version 25. ANN is an innovative method for predicting labor productivity, which can solve non-linear problems and learn from examples and training data. The model used 67.8% for training, 23.7% for testing and 8.5% for validation. The error between actual productivity and estimated productivity was computed using Mean Square Error (MSE) which was 0.0019 which verified that the estimated production rate was within an acceptable range. RII evaluation reveals that the top ten factors are worker skill, worker experience, instructions to workers (communications), material availability, worker health, contractor experience, incomplete drawings, planning and scheduling, and site management respectively. The ANN model also predicts similar influential factors among the top ten as worker experience, contractor experience, worker skill, site management, and alterations in drawing/specifications indicating the close alignment between two. vii The average of actual and predicted productivity values is 0.292 and 0.297 m3 /man-day respectively which are both more than the standard DUDBC productivity value 0.242 m3 /man-day. Also, the developed model successfully predicted the productivity value on the given data set matching the site conditions. From the Key Informant Interview (KII) conducted with experts in the building construction field skill of workers, experience of workers, timely payment, material availability, supervision, availability of workers, instruction to workers / multiple instruction, site management / management of resources are determined to be the main impacting factors.

Keywords

Building Construction, Labor Productivity of Brickwork, Artificial Neural Network (ANN) , Relative Importance Index (RII)

Back to list