Review article

Regression assessment of labor cost standards based on neural network simulation

UDC: 

658.5:624.05

DOI: 

10.23968/1999-5571-2020-17-3-127-133

Pages: 

127-133

Annotation: 

Construction scheduling is based on data on the duration of works, which are directly related to the values of labor costs. This ensures obtaining of the most accurate results of scheduling implemented both at the stage of construction design and in the process of its operational management. In the Russian Federation and in other leading foreign countries, there have been developed databases for standardized labor costs in the field of construction. However, many developing countries, for example, the Republic of Iraq, are not provided with their own standards for labor costs, and therefore, they are forced to use either the standards of other countries or determine the duration of work with the help of expert estimates, which leads to a significant error in the construction planning schedule. To create databases on labor standards, various methods are used that are either based on timing measurements, or on physiological methods for measuring anthropological parameters associated with the energy costs of workers. Since the use of these approaches requires significant costs and time, a technique is proposed as an acceptable alternative. This technique is based on the collection of statistical data included in the executive documentation for construction, expert data that determine the feasibility of regulatory conditions for the performance of work, and regression processing of the final results based on the use of modeling capabilities of artificial neural networks.

Authors: 

Bolotin S. A. Saint Petersburg State University of Architecture and Civil Engineering St. Petersburg, Russia

Khoshnaw Yousif Babakr Hussein Saint Petersburg State University of Architecture and Civil Engineering St. Petersburg, Russia

Bohan Haitham Saint Petersburg State University of Architecture and Civil Engineering St. Petersburg, Russia

Links to authors' articles: 

Journal Issue