En jämförande studie av heuristikbaserade sammanslagningsstrategier i trafikkoordinering

Efficient traffic merging is critical for minimizing congestion, particularly at lane reduction zones and highway on-ramps. This study investigates the effectiveness of four heuristic-based merging strategies: Time-Based Traffic Light, Priority-Based Traffic Light, Late Merging (Zipper Merge), and Auction-Based Right-of-Way Allocation, using a custom simulation framework. Each heuristic is evaluated across varying traffic densities and assessed using key performance metrics: throughput, average passing time, average stationary duration, stop count, fairness, and traffic stability. Results show that the Late Merging strategy consistently minimizes stops and stationary durations, supporting smoother flow and greater stability. Auction-Based Right-of-Way Allocation achieves the highest throughput under heavy traffic, albeit with increased stationary times. The Priority-Based Traffic Light strategy offers only modest improvements over the Time-Based approach, with both showing limited adaptability in dynamic traffic conditions. Although these heuristics perform well in controlled simulations with perfect compliance, their real-world applicability is subject to behavioral variability and infrastructure limitations.