This report deals with questions concerning validation of the Swedish national goods transportation model, Samgods. To some degree, also more general issues of validation are discussed.
Go to the publication
When developing transportation models it is essential to have independent data available that can be used for comparison, validation and calibration. The quality and availability of such data are obvious issues. It is a big advantage if these data cover many different aspects of the model results and are collected on a regular basis.
In the report, a survey of data sources useful for validation purposes is presented. For each source of data, quality and availability is discussed. For statistical data, the method for collecting the data is described in some detail as well as uncertainties and access to the data that may be restricted due to privacy regulations.
There are reasons to believe that in the future new technologies will provide us with new types of data useful for validation or model estimations. A special section in the report is devoted to investigating if present day ITS systems can provide data for validation purposes.
Once the compilation of validation data is completed, many problems remain concerning the validation. In particular, there is a matching problem between the model output data and validation data. The many different table formats and aggregation levels for the data add to the complexity of this problem. A similar, but easier, problem is to compare output data from different versions of Samgods. Within this project a computer program has been developed that can be used for matching such kinds of datasets. The program, as well as a user manual for it, is included in the report. The possibility to handle some of the issues concerning confidential data, by using the program, is briefly touched upon.