In the English-speaking world (UK, USA, Australia, NZ), data on the use of public transport services is generally classified according to the highest-ranking transport system. This data is used to validate demand models and assignment results. For traffic modelers, this means that the assignment and the calculation of skims must be based on a set of connections that uses a transport supply that is mandatory on at least part of a connection.
The selection of the mandatory supply is based on time profiles. A time profile that is mandatory for use must have an (indirect) attribute not equal to 0. Referencing in the assignment procedure is carried out via the parameters in the assignment procedure. Due to the dependence on time profiles, the use of supply of the type PuT-Aux cannot be made mandatory.
The procedure also allows the definition of two disjoint transport supply sets, which are logically linked with AND, NOT AND, or EITHER OR, depending on the setting. This can be used, for example, to model situations in which two competing companies do not have a common ticket or trip planning, so that there is usually no switching between the providers during a journey.
The mandatory use of a transport supply is implemented in the structure of the search graph and is therefore independent of the passengers' level of information or the parameters of the choice model. The functionality can be combined with the fares. However, this should be used with caution, as the search graph becomes significantly larger and the duration of the procedure increases.
The addition of the assignment procedure offers the possibility of realistically estimating paths from survey matrices and calibrating corresponding demand models using realistic skims. The computing time and memory requirements increase with the extent of the restrictions on the transport supply. If the mandatory transport supply does not match the demand, unrealistic connections are created, sometimes with very long detours.
