Wykorzystanie sztucznych sieci neuronowych do oceny możliwości wystąpienia opóźnień w realizacji kontraktów budowlanych

Hubert Jerzy Anysz

Abstract

Forecasting completion deadlines of road sections constructions is shown as one of the risk management instruments, allowing to limit the negative impact of possible delays which negative financial, material and other consequences affect as investors as contractors, and also the general public. The dissertation focuses on creating an effective instrument which could be successfully applied in the prognosis of not meeting the planned deadlines concerning the construction of the sections of express roads and highways in Poland. The construction of such instrument required identification of possible causes of delays. The previously mentioned task is carried out based on analysis of the questionnaire directed to engineers involved in construction processes – members of the Chamber of construction engineers and based on analysis of the respective studies concerning similar Polish and international research. The list of possible causes of delays is supplemented with the author’s own analysis, regarding the influence of the economic factors, selected contract clauses, contractors’ features as well as features of the constructed objects. Additionally, the analysis is made of important – for the author – factors that may cause delays, and factors which could be defined before choosing a contractor are selected. The data for the studies refers to the contracts for express roads and highways construction in Poland, ordered by Generalna Dyrekcja Dróg Krajowych i Autostrad (the public investor for national roads and highways in Poland) completed between 2009 and 2013. Artificial Neural Networks is chosen as the predicting instrument and Matlab with Neural Networks Toolbox as the software. Topology and other parameters of ANN are optimized in order to achieve the most accurate predictions. Though there is a variety of methods applied in errors’ measurement, the author creates his own sectional measure of accuracy predictions. That allows to choose the only one Artificial Neural Network giving the most accurate predictions and effect the final calculations. The simulation of bidding procedure concerning construction of a road section is created. The author shows the possibility of applying the prognosis obtained with the help of ANN as one of the criterion for the assessment of offers (tenders) submitted and as a premonitory prognosis enabling an investor possible countercheck of the planned deadline, previously set for the construction of the road’s section. Finally, the result achieved are summarized, and the final conclusions are presented. The dissertation inspired the author for further analysis and research, which foundations are described at the end.
Diploma typeDoctor of Philosophy
Author Hubert Jerzy Anysz (FCE / ICE)
Hubert Jerzy Anysz,,
- The Institute of Civil Engineering
Title in PolishWykorzystanie sztucznych sieci neuronowych do oceny możliwości wystąpienia opóźnień w realizacji kontraktów budowlanych
Languagepl polski
Certifying UnitFaculty of Civil Engineering (FCE)
Disciplineconstruction / (technology domain) / (technological sciences)
Start date13-03-2013
Defense Date18-10-2017
End date25-10-2017
Supervisor Artur Zbiciak (FCE / IRB)
Artur Zbiciak,,
- The Institute of Roads and Bridges

Nabi Ibadov (FCE / ICE)
Nabi Ibadov,,
- The Institute of Civil Engineering

Pages280
Keywords in Englishartificial neural networks, ANN, MLP, predictions, delays, completion dates, construction site
Abstract in EnglishForecasting completion deadlines of road sections constructions is shown as one of the risk management instruments, allowing to limit the negative impact of possible delays which negative financial, material and other consequences affect as investors as contractors, and also the general public. The dissertation focuses on creating an effective instrument which could be successfully applied in the prognosis of not meeting the planned deadlines concerning the construction of the sections of express roads and highways in Poland. The construction of such instrument required identification of possible causes of delays. The previously mentioned task is carried out based on analysis of the questionnaire directed to engineers involved in construction processes – members of the Chamber of construction engineers and based on analysis of the respective studies concerning similar Polish and international research. The list of possible causes of delays is supplemented with the author’s own analysis, regarding the influence of the economic factors, selected contract clauses, contractors’ features as well as features of the constructed objects. Additionally, the analysis is made of important – for the author – factors that may cause delays, and factors which could be defined before choosing a contractor are selected. The data for the studies refers to the contracts for express roads and highways construction in Poland, ordered by Generalna Dyrekcja Dróg Krajowych i Autostrad (the public investor for national roads and highways in Poland) completed between 2009 and 2013. Artificial Neural Networks is chosen as the predicting instrument and Matlab with Neural Networks Toolbox as the software. Topology and other parameters of ANN are optimized in order to achieve the most accurate predictions. Though there is a variety of methods applied in errors’ measurement, the author creates his own sectional measure of accuracy predictions. That allows to choose the only one Artificial Neural Network giving the most accurate predictions and effect the final calculations. The simulation of bidding procedure concerning construction of a road section is created. The author shows the possibility of applying the prognosis obtained with the help of ANN as one of the criterion for the assessment of offers (tenders) submitted and as a premonitory prognosis enabling an investor possible countercheck of the planned deadline, previously set for the construction of the road’s section. Finally, the result achieved are summarized, and the final conclusions are presented. The dissertation inspired the author for further analysis and research, which foundations are described at the end.
Thesis file
h_anysz_rozprawa_2017.pdf 5.36 MB
Citation count*8 (2020-08-24)

Get link to the record

Back
Confirmation
Are you sure?