PRIORITIZATION OF EVENTFUL ROADS CORRECTION USING ARTIFICIAL NEURAL NETWORKS

Eventful sections prioritization methods usually have been based on single-criterion methods. Therefore, with respect to a separate and distinct criterion, identifying and prioritization of different sections have been done. Due to budget deficiency for immunization actions, prioritization non-immune sections are of great importance. In this regard, several different methods were presented by experts that were based on a specific basis. Among these, we can point to two general structures based on economic and technical studies. On the other hand due to weakness in accidents data (incomplete, incorrect or unused data), It is better to identify and prioritize the proposed methods, evaluated as far as possible without regard to the accidents data, here the new and innovative methods have special significance. In this study which is new in the field of road immune in the country, we use powerful device named multilayer Artificial Neural Networks for predicting prioritization of eventful MAZANDARAN road networks modifying. And tried to investigate the advantages of using neural networks, considering the history of accidents and the costs of their modifying, achieving method to optimal prioritization should be analyzed.