استخدام نموذج الشبكات العصبية الاصطناعية في تقدير تكاليف مشاريع الطرق في محافظة الانبار غرب العراق

Abstract EN

The cost estimate in road projects in Anbar province and with high accuracy and in the stages of development and development is an important and crucial role in planning in the preparation of feasibility studies, technical, economic , and improve the level of municipal services , and in any case show many difficulties when estimating or calculate the costs of road construction during the preparation phase and planning for those projects, where facing departments planning and management are major problems in the lack of preliminary information and the lack of database and determine the costs and data inaccurate in addition to the lack of scientific methods to estimate the costs and therefore the way networks troubled industrial used in this study reduces the problems predict costs and is widely used to address those the problems this way a new methodology in building models , analysis and evaluation of data , forecasting and control.

In this study, has been using artificial neural networks to create a model to calculate the cost of road construction in Anbar province.

I have been using a database obtained from the Directorate of ways Anbar have included a total of (152) by major and minor .

The model inputs was Klaty : lengths of roads (km) , dirt business volume (thousand cubic meters) and the implementation period (days) The output of the model was the cost of road construction (million Iraqi dinars).

In this study, the use of multi-layered networks retrograde spread of technology Khtalnmzjh construction costs .

Has been studying many of the cases related to the construction of Artificial Neural Networks , including network architecture and internal factors have and how Taterhaaly performance models Artificial Neural Networks, and information networks have been developed information about the relative importance of the factors affecting the calculation of costs and put the equation calculation to calculate those costs.

I've found that the Artificial Neural Networks have the ability to predict the cost of road construction degree good accuracy as the models that have been built to study the effect of the internal factors of the networks on their performance showed that the performance of the network is sensitive to the number of nodes in the layer hidden, learning rate , equations of transport and variable momentum.

Has shown a sensitivity analysis to model the costs that the business volume and lengths of dirt roads have a bigger role in determining the costs followed in importance for implementation and the relative importance of (0.98 , 0.98 , 0.77), respectively.

The equation of industrial neural network model which is found in this study to calculate the costs were complicated calculations that were written and programmed to facilitate and speed up the calculations