Application of biological algorithms in standardization of construction work

In the Czech Republic, established methods in work standardization are based solely on mathematical statistics. The author suggests using neural networks models in standardization of construction works or for other predictions in preconstruction phases. Such models are able to predict e.g. the time of implementation of the project and price of implemented projects based on the learned context (experience). Assuming that sufficient data will be available thanks to technological progress and the implementation of Building Information Modeling (BIM) and further research will be done in the field of neural network optimization, the suggested procedures should lead to the optimal predictive model. Preparatory processes will be automated and there will be more time left for qualified professional work. The main objective of the thesis the designing and testing models created using neural networks has been achieved through three examples which based on theoretical and real data predicted the selected outputs and examined the relationship between the amount and detail of input data and quality of neural network model and encompassed methods determining the parameters that have an unavoidable effect on the resulting values.