Hybrid Rule-Based and Reinforcement Learning for Urban Signal Control in Developing Cities: A Systematic Literature Review and Practice Recommendations for Indonesia

Hybrid rule-based and reinforcement-learning (RL) signal control is gaining traction for urban coordination by pairing interpretable cycles, splits, and offsets with adaptive, data-driven updates. However, systematic evidence on their architectures, safeguards, and deployment prerequisites remains scarce, motivating this review that maps current hybrid controller designs under corridor coordination. Searches across major databases and arXiv (2000–2025) followed PRISMA guidance; screening is reported in the flow diagram. Eighteen studies were included, nine with quantitative comparisons, spanning simulation and early field pilots. Designs commonly use rule shields, action masking, and bounded adjustments of offsets or splits; effectiveness is assessed via arrivals on green, Purdue Coordination diagrams, delay, and travel time. Across the 18 studies, the majority reported improvements in arrivals on green, delay, travel time, or related coordination metrics compared to fixed-time or actuated baselines, while only a few showed neutral or mixed effects and very few indicated deterioration. These results indicate that hybrid safeguards are generally associated with positive operational gains, especially under heterogeneous traffic conditions. Evidence specific to Indonesia remains limited; this review addresses that gap and offers guidance transferable to other developing-country contexts with similar sensing, connectivity, and institutional constraints. Practical guidance synthesizes sensing choices and fallbacks, controller interfaces, audit trails, and safety interlocks into a deployment checklist, with a staged roadmap for corridor roll-outs. This paper is not only a systematic review but also develops a practice-oriented framework tailored to Indonesian corridors, ensuring that evidence synthesis and practical recommendations are clearly distinguished.