Intelligent Meta-Heuristic-Based Optimization of Traffic Light Timing Using Artificial Intelligence Techniques

This research examines worldwide concerns over traffic congestion, encompassing aspects such as security, parking, pollution, and congestion. It specifically emphasizes the importance of implementing appropriate traffic light timing as a means to mitigate these issues. The research utilized a dataset from Montreal and partitioned the simulated area into various zones in order to determine congestion levels for each individual zone. A range of prediction algorithms has been employed, such as Long Short-Term Memory (LSTM), Decision Tree (DT), Recurrent Neural Network (RNN), Auto-Regressive Integrated Moving Average (ARIMA), and Seasonal Auto-Regressive Integrated Moving Average (SARIMA), to predict congestion levels at each traffic light. This information was used in a mathematical formulation to minimize the average waiting time for vehicles inside the road network. Many meta-heuristics were analyzed and compared, with the introduction of an Enhanced Bat Algorithm (EBAT) suggested for addressing the traffic signal optimization problem. Three distinct scenarios are described: fixed (with a constant green timing of 40 s), dynamic (where the timing changes in real-time based on the current level of congestion), and adaptive (which involves predicting congestion ahead of time). The scenarios are studied with low and high congestion scenarios in the road network. The Enhanced Bat Algorithm (EBAT) is introduced as a solution to optimize traffic signal timing. It enhances the original Bat algorithm by incorporating adaptive parameter tuning and guided exploration techniques that are informed by predicted congestion levels. The EBAT algorithm provides a more effective treatment for congestion problems by decreasing travel time, enhancing vehicle throughput, and minimizing pollutant emissions.