Intelligence series: How will algorithms provide solutions humans can't see?
Rail traffic managers could use network optimisation algorithms to prevent delays from cascading throughout the network. For example, network optimisation algorithms can be used for route-planning to compute solutions for different train demand scenarios, thereby distributing traffic more efficiently. Rolling stock rotation planning could benefit from optimisation algorithms as they can reduce the number of deadhead trips and the cost of additional turn around trips, increasing operational efficiency. Although Train Management Systems (TMSs) estimate train movements and conflict identification, few make re-scheduling and re-routing decisions. Consequently, network optimisation algorithms can be integrated into TMSs to assist real-time train dispatching and compensate for unexpected events such as train delays, network failures and cancellations. Furthermore, the technology can be used preventatively for delay time prediction. Systems can use machine learning techniques to predict the length of journey delays and deliver updates, allowing passengers to receive more accurate delay information and an enhanced passenger experience. Wireless sensor networks (for remote conditioning monitoring) can be designed using algorithms to maximise their lifetime and reduce the operational cost. Network optimisation algorithms can deliver key capabilities of the RTS, including providing more value from data, timing services to the second, running trains closer together and minimising disruption to trains.