Estimation of cost uncertainties in construction projects for rock and earthfill dams using artificial intelligence techniques

There are approximately thirty different work items for the construction of a dam structure. Due to variables such as the type of the project, the amount and type of filling materials used, too many uncertainties occur during planning and implementation of projects. These uncertainties prolong evaluation of the approximate cost estimates. Uncertainties in the approximate cost also disadvantage the bidders bidding for the project, contractor companies that take the job may suffer losses due to incorrect price offers that they may give, and jobs may be interrupted as well. It is seen by examining the literature that the approximate cost can be calculated with high accuracy with artificial intelligence and mathematical models to estimate it in a short time. In this study, 114 dam and pond projects, whose tender were made and the results were announced, between 2011 and 2020 within the scope of the Public Procurement Law No. 4734 (KIK) were examined and the work items that were important for the approximate cost were determined. To correlate approximate costs across all projects, prices were fixed to the December 2020 prices. By using these factors different approaches were created and the generated approaches were tested with regression models. The weight ratios of the regression models were calculated using the gray wolf algorithm. As a result of this study; it is found that dam construction through the Public Procurement Law, where there is little pre-tender data and the approximate cost needs to be determined quickly, it can be determined with high accuracy rates with mathematical models.