Akademik Veri Yönetim Sistemi
MEKON 2025 Uluslararası Katılımlı Öğrenci Konferansı , İstanbul, Türkiye, 26 - 27 Haziran 2025, ss.5-8, (Tam Metin Bildiri)
Traffic congestion is a pervasive global challenge, leading to significant economic losses, environmental degradation, and reduced quality of life. Traditional traffic signal control (TSC) strategies, such as fixed-time and basic actuated systems, often fail to adapt effectively to dynamic and complex urban traffic conditions. This paper addresses an adaptive approach to traffic signal control by developing and implementing a prototype system based on Monte Carlo Tree Search (MCTS). Utilizing the Simulation of Urban MObility (SUMO) environment integrated via the Traffic Control Interface (TraCI), the proposed MCTS controller aims to optimize signal timings by intelligently exploring potential future traffic scenarios. The core components of the MCTS framework, including state representation, action selection and reward function are tailored for the traffic control domain. This study focuses on demonstrating the feasibility and operational principles of an MCTS-based controller for a single intersection. Preliminary observations from the SUMO simulation indicate the prototype's capability to make adaptive decisions in response to fluctuating traffic demands.