Travels with My App: Computational Intelligence for Individualised Dynamic Rail Journey Planning

This presentation describes ongoing research for the Transport iNet funded project Evolutionary Computation for Optimised Rail Travel (EsCORT), which is running from 11th November 2013 to 13th December 2014.

The recently completed first phase of the project was a literature review which considered what computational intelligence might offer to dynamic, individualised, passenger rail journey planning. Of particular interest was whether it is feasible to create an app to dynamically assist in selecting and updating rail journey itineraries, and if so what improvements this would bring relative to a dynamically updated timetable. The study was concerned specifically with station to station railway journeys within the British Railway System.

There was found to be nothing in the published literature referring to an app that plans passenger rail journeys, then responds to unforeseen disruptions by alerting the passenger and offering an adjusted itinerary. The literature review concluded that the development of such an app would therefore be timely, filling a gap in the market.

The project is now at its mid-point according to the calendar. However only one sixth of the days funded have been worked; the remaining five sixths will be devoted to investigating the journey planning algorithms that would drive the app. To begin with, the extent to which solutions may be arrived at deductively through mathematical modelling is being explored. The relevant field of mathematics here is graph theory.

Complexity theory is also highly pertinent; it is anticipated that any mathematical model produced will be far too computationally complex to be incorporated into usable software. This negative result would justify a computational intelligence approach. The three main forms of computational intelligence, each in its own way inspired by nature, are artificial neural networks, evolutionary computation, and fuzzy logic; each paradigm’s potential for contributing to passenger journey planning will be evaluated.