The subject of the article is to present the possibility of applying artificial neural network in analysis of observational data on the issue of simulation concrete supplies in construction industry. Neural networks allows a quick and efficient way to model phenomena at different levels of complexity. They can be used both in processes of well and poorly structured problems, knowledge of which is limited. The advantages of neural networks were transferred to the possibilities of defining probability distributions of individual tasks in cyclical building processes. The article presents the problem of concrete supplies for construction site, its elements and parameters which describe the process. The use of simulation in decision-making system was also demonstrated. One of the stages of creating simulation models is the adoption and verification of probability distributions that describe the elements of the system. In order to learn and predict the characteristics of probability distributions neural networks were used, that enable in the efficient and effective manner the implementation of further stages of simulation modeling. The article presents the results of applying the neural network and approach that enables their use.
Gajzler, M., & Konczak, A. (2015). The Possibility of Using Neural Networks in Data Analysis Connected with Observation in the Construction Process Simulation. In Procedia Engineering (Vol. 122, pp. 228–234). Elsevier Ltd. https://doi.org/10.1016/j.proeng.2015.10.029