Article ID: | iaor20032858 |
Country: | United States |
Volume: | 18 |
Issue: | 4 |
Start Page Number: | 303 |
End Page Number: | 329 |
Publication Date: | Jan 2001 |
Journal: | Civil Engineering and Environmental Systems |
Authors: | Chua D.K.H., Kog Y.C., Loh P.A.R.K. |
Keywords: | engineering |
Ensuring satisfactory budget and schedule performance are two major challenges for construction projects. On this issue, predictive models for budget and schedule performances can provide assistance in the appropriate allocation of project management resources. Two neural network models for construction budget and schedule performances using fuzzy data have been developed in the present study. These models consist of eight and five key determinants of project outcome, respectively. A combined fuzzy index (CFI) approach is introduced for data representation. The CFI for an input or output measurement can be derived using the fuzzy number membership degree concept. This approach permits a gradual change of scale value in the classification. Several definitions to the fuzzy numbers are experimented. The results reveal that this approach is a feasible alternative for neural network implementations dealing with quantitative measurements. Examples of using the models for guidance in project management are presented. These include the effect of amount of design completed before construction starts on budget performance, and the effect of amount of time devoted by project manager on schedule performance. The trade-off effect between two key determinants on project outcome can also be studied using the models.