Statybos projektų valdymo efektyvumo modeliavimas, taikant dirbtinius neuroninius tinklus

LTStraipsnyje siūloma nauja metodika statybos projektų valdymo efektyvumui įvertinti. Pateikiama statybos projektų valdymo veiksnių analizė ir dirbtinių neuroninių tinklų (DNT) metodologijos taikymas statybos projektų išlaidų kitimo prognozės modeliavimui. Siūlomasis metodas leidžia nustatyti statybos projekto esminius valdymo veikslius, projekto išlaidų kitimo ribas ir realiai įvertinti projekto valdymo riziką. Esminių projektų valdymo veiksnių identifikavimas DNT pagrindu leidžia nukreipti ir sukoncentruoti įmonių ir projektų vadovų dėmesį į svarbiausius valdymo sistemos klausimus ir sumažinti projektų tikslų įgyvendinimo riziką. [Iš leidinio]Reikšminiai žodžiai: Dirbtiniai neuroniniai tinklai; Statybos projektų valdymas; Statybos projektų valdymas, dirbtiniai neuroniniai Tinklai; Construction project management, artificial neural networks.

ENConstruction projects are delivered under conditions of risk in the competitive market environment There are external risks (economic, political, financial and environmental) and internal risks based on project management issues, i.e. projects manager's and his team competency, experience, strategic and tactic decisions made during construction project delivery. The opportunity to improve organizational performance through more effective project management could provide substantial savings for construction management company. The paper presents a new methodology for construction project management effectiveness modelling by applying artificial neural net works. Artificial neural networks are defined as a new information processing technique of artificial intelligence inspired by the biological brain model. The approach of artificial neural networks allows the construction projects management effectiveness model to be built and to determine the key determinants from a host of possible management factors that affect project effectiveness in terms of construction cost variation. The historical data of project performance has been used to build the neural network model. The present study is based on a set of data obtained in a questionnaire survey on construction project management effectiveness factors from construction management organisations in Lithuania and the US. A list of construction management factors was collected according to the results of past research and opinion of experienced construction management practitioners.Construction project management effectiveness modelling by applying neural networks consists of the following stages: selection of the variables of the construction project management effectiveness neural network model (CPMEM); selection and preparation of training data for the CPMEM; designing and training the construction project management effectiveness neural network; evaluation of the importance of a particular input factor to the CPMEM output by applying a sensitivity analysis technique; identification of the key construction project management effectiveness factors and modification of the CPMEM; determining the validation range of the CPMEM practical applications. [...] The application algorithm of the construction project management effectiveness model was developed. By applying the construction project management effectiveness neural network model, managers of construction company can indicate how much importance each factor has for a particular project outcome, find the best possible arrangement of construction management effectiveness factors and examine the construction cost variation tendencies. The established neural network model can be used during the competitive bidding process to evaluate management risk of a construction project and predict construction cost variation. The model allows the construction project managers to focus on the key success factors and reduce the level of construction management risk. [From the publication]