FASTA has been developed for assessing timetable structures in strongly connected networks. It is built for "sketch simulation", as it does not aim to produce detailed results but rather to assist in identifying major discrepancies between network and timetable, at early planning stages. In the new planning paradigm (operations first and, then, infrastructure), long term vision starts by defining a timetable in only its basic structure, leaving spare capacity for allocation in a later stage (a

more »

... ear, a month, even a day before operation). Moreover, in long term timetable planning, many alternatives are drawn. To assess reliability of timetables designed 20 or 30 years ahead, there is no need (and no possibility) to use detailed simulation, which requires detailed definition of infrastructure and signalling. To apprehend the basic behaviour of a timetable under multiple assumptions, planners need at this stage to run a huge number of simulations fast and efficiently. Here comes the utility of sketch simulations and tools, such as FASTA, able to perform them efficiently. The main question during development is to set the right balance between modelling simplicity (unavoidable due to uncertainties on far-future infrastructure and timetable) and accuracy of the results. The trade-off problem between simplicity and accuracy bears no single solution. Therefore, the initial options that have been taken for FASTA some twenty years ago had to be revisited. Developers had to reassess the consequences of simplification and to extend the concept, in order to build a new package, based on a sensibly modified balance, by trading off simplicity for better accuracy. The paper presents the basic concepts of the simulation package, the initial trade-off between simplicity and accuracy and its justification, the reasons for changing the initial solution, and the latest developments of the FASTA package based on a changed scope in planning needs.