A Literature Review on the Integration of Game Theory and Artificial Intelligence in Traffic Signal Control

In the twenty-first century, artificial intelligence (AI) has been one of the most promising breakthroughs in computer science, revolutionizing intelligent transportation systems (ITS). Among the pressing challenges within ITS is traffic congestion. Congestion undermines urban mobility, economic productivity, and environmental sustainability. Traditional TSC systems struggle to adapt to the dynamic and multiagent nature of real-world traffic. To address this, researchers have increasingly turned to AI and game theory (GT). The two fields offer powerful tools for modeling and optimizing strategic decision-making in complex environments. While AI enables systems to learn from data and respond adaptively, GT provides a mathematical framework for reasoning about the interactions and incentives of multiple agents in signal control. Studies have shown how each field can successfully optimize TSC operation. Furthermore, studies have shown that integrating AI and GT has been successful in domains such as adaptive traffic signal control. This integration has been relevant when decentralized agents must coordinate and when conflicting objectives require substantial computational resources to be fully optimized. Moreover, advancements in multiagent learning, mechanism design, and behavioral modeling have further deepened the synergy between these fields. This paper presents a literature review examining the most recent (2018–2025) research progress at the intersection of GT and AI in traffic signal control. We highlight key application areas at both isolated and multi-intersection levels. We also identify open research challenges/limitations and propose potential future research areas for practical use.