Analysis of route selection preferences in urban transportation with artificial intelligence methods (Balikesir case)

With the impact of urbanization and population growth, the problems encountered in transportation systems are becoming increasingly complex, and route choice stands out as a significant factor directly affecting road users' travel experience. Route preferences are shaped by the interaction of numerous variables such as distance, time cost, traffic conditions, safety, environmental factors, and road quality. However, the uncertainty in individual preferences complicates modeling these preferences. This complicates effective traffic planning decisions. This study aims to examine the preference behaviors of road users for ten alternative routes located between Sanayi 2. Kapısı and Yeni İzmir Caddesi within the borders of Karesi and Altıeylül, the central districts of Balıkesir Province, using survey data and to model these behaviors using a fuzzy logic approach. Because the survey data is based on real user experiences, the validity and reliability of the model are enhanced. The fuzzy logic approach and artificial neural networks effectively reflect the uncertainty, imprecision, and insensitivity in preference behaviors, providing flexible and realistic modeling opportunities. The developed model predicts user preferences more successfully than traditional deterministic methods and offers highly applicable results in traffic planning and management processes. The study contributes to the development of transportation policies, improved traffic management, and the creation of sustainable transportation strategies. It also forms the basis for strategic planning aimed at understanding private vehicle users and public transportation preferences. This study demonstrates the potential of artificial intelligence methods in solving complex, multivariate, and uncertain transportation problems. The study reveals that, specifically in Balıkesir Province, distance, time, and traffic density are the most important determinants of route choices. The findings demonstrate that, compared to traditional deterministic models, fuzzy logic and artificial neural networks can develop reliable and applicable models even with small sample sizes.This study demonstrates the potential of artificial intelligence methods in solving complex, multivariate, and uncertain transportation problems. Compared to traditional deterministic models, the combined use of fuzzy logic and artificial neural networks has been shown to more effectively manage uncertainties in real-world data. The results confirm the ability of these two methods to predict subjective factors such as user preferences with high accuracy. KEYWORDS: Optimization and its techniques, transportation planning, transportation and traffic