Dynamic Segmentation Method for Identification of Accident-Prone Road Segments using Genetic Algorithm

This paper presents heuristic methods for identification of accident-pronesegments in using Genetic Algorithm in with dynamic segmentation method.Physical and mental injuries as well as large social and economic costs make thegovernments pay attention to the road safety carefully. This study focus onfinding the location of accidents that are occurring on the road. In fact, forlimiting the large budget to repair and maintenance the some parts of the road toimprove the safety of the road, we should find the areas and segments that havemore accidents on them. So, it's so important to identify the segments with highnumber of accidents and their lengths in order to optimize road safety with thelimited budget. In this study, first we suggest a novel chromosome codingmethod and a fitness function can work with Genetic Algorithm. Then wecompare the proposed method with the other fixed segmentation methods. Theproposed method is as a dynamic segmentation method that has huge advantagescompared to static segmentation methods. This method has great improvementon the variance of the accident density, which has more than two times highervariance compared to the fixed conventional methods.

کلیدواژه ها:

Dynamic Segmentation ، Genetic Algorithm ، Identification of Accident-Prone Segments