Speakers
Description
Major events in Riyadh can generate sudden passenger surges that lead to severe platform crowding, long waiting times, and slow crowd dispersal at key interchange stations such as the King Abdullah Financial District (KAFD) Metro Station. Existing metro systems support passenger monitoring and demand forecasting; however, operators have limited ability to evaluate alternative interventions before implementing them. This paper presents an AnyLogic-based predictive and prescriptive decision-support system that enables metro operators to simulate, compare, and select operational strategies for managing event-based congestion. The model represents passenger arrivals, platform queues, train boarding, train movement, capacity constraints, and passenger exit flows. Four scenarios were evaluated using consistent performance indicators: a six-minute baseline headway, a four-minute improved service, a three-minute empty relief train strategy, and a two-minute train-reallocation strategy. The results show that train frequency is the strongest control variable affecting platform recovery. Compared with the baseline, the two-minute train-reallocation scenario reduced average passenger waiting time by 95.2% and crowd dispersal time by 65.5%. The three-minute empty relief train scenario provided a practical compromise, reducing waiting time by 71.9% and dispersal time by 49.3% with lower operational intensity. This case demonstrates how simulation-based decision support can provide a safe virtual environment for testing operational interventions and support proactive, evidence-based decisions to improve passenger safety, service efficiency, and resilience during major events.
| Topic Area | Smart Operations and Maintenance |
|---|