Paper Title
An Application Of Smeed And Andreassen Accident Models For The City Of Ankara By Differential Evolution Algortihm

Abstract
Traffic accident prediction models are commonly used for traffic safety studies. Because of their successful performances, heuristics algorithms have recently been used in traffic studies. In this study, for the city of Ankara, two analytical models to predict the number of accidents, fatalities and injuries were developed by employing Differential Evaluation (DE) Algorithm. In the development of the models, the forms of Smeed and Andreassen accident prediction models which are widely used in the literature were utilized. The number of vehicles (N), fatalities (D), injuries (I), accidents (C) and population (P) were taken as the model parameters. 16 years of data taken randomly from 21 years historical data were used to develop the analytical models, and the rest of them a-five year data were employed for testing the developed models. The performances of the developed models were statistically evaluated in terms of error criterias which are root mean square errors (RMSE) and mean absolute errors (MAE). Two scenarios was considered to investigate the performances of the developed models for the future estimates, In the first scenario, the average number of vehicles per capita is assumed to reach 0.40 in 2023 and in the second scenario, this ratio is supposed to reach 0.60. According to the results obtained from the scenarios, some suggestions were presented for traffic safety applications. Keywords- Accident prediction models, Smeed Model, Andreassen model, Ankara