COMPARISON OF DIFFERENT APPROACHES IN TRAFFIC FORECASTING MODELS FOR THE D-200 HIGHWAY IN TURKEY

DOGAN, Erdem; KORKMAZ, Ersin; AKGUNGOR, Ali Payidar · 2018 · Crossref

DOI: 10.20858/sjsutst.2018.99.3

archive: archived pipeline: cataloged verified

Get this paper ↗ (DOI — opens at the source; we link to it, we don't host it)

Summary

This study addresses the need for efficient short-term traffic forecasting to improve traffic management and control on the D-200 highway in Turkey. The authors developed and compared three distinct modeling approaches: Seasonal Autoregressive Integrated Moving Average (SARIMA), Differential Evolution (DE), and Artificial Bee Colony (ABC) algorithms. While SARIMA is a established statistical method, the application of DE and ABC algorithms for traffic estimation represents a novel contribution, as these heuristic optimization techniques had previously been used for signal optimization but not for direct traffic flow prediction. The research utilized 4,512 traffic flow data points collected from NC-350 counting devices on a two-lane section of the D-200 highway. Data were recorded at 15-minute intervals, with the right lane primarily used by heavy vehicles and the left lane by lighter traffic. The dataset was split into 80% for training and 20% for testing. The SARIMA model was developed using the Box-Jenkins method, identifying the optimal structure as SARIMA (1,0,1) (0,1,1)672 after stabilizing the non-stationary time series. The DE and ABC models were constructed using linear, semi-quadratic, and power forms, utilizing traffic data from the preceding hour as input variables. The DE algorithm employed a population size of 30 and a mutation strategy of DE/best/1/exp, while the ABC algorithm used a colony of 50 bees. Model performance was evaluated using Mean Absolute Error (MAE), Mean Absolute Percentage Error (MAPE), Root Mean Square Error (RMSE), and the coefficient of determination (R²). The results indicated that all models provided consistent and useful forecasts. The SARIMA model achieved R² values of approximately 0.89 for the right lane and 0.85 for the left lane. The DE and ABC models generally yielded higher R² values, with the right lane models reaching approximately 0.91–0.92 and the left lane models around 0.87–0.88. Specifically, the ABC algorithm demonstrated the lowest error rates across MAE and RMSE metrics compared to the SARIMA and DE approaches. For instance, the ABC model’s linear form showed superior performance in minimizing prediction errors. The study concluded that the ABC algorithm is the most effective approach among those tested for short-term traffic forecasting on this highway. The authors suggest that the ABC model is appropriate for application on other highways in Turkey, highlighting its potential for integration into intelligent transport systems to enhance traffic control efficiency.

Provenance

The full processing record for this entry. Every stage of this paper's journey through the pipeline is logged — what ran, with which tool and model, how many attempts it took, and when it last completed.

Summary generated by qwen3.6-27b-prismaquant on 2026-06-26; verification: verified.

Topics

Ranked by relevance to this paper. Hover a topic for its definition.

• mental model of traffic