I am doing a comparison of different algebraic modeling langues (AMPL, AIMMS, GAMS, Pyomo) in both theoretical and practical terms. As a practical experiment I am trying to measure problem model instance creation time. I.e. taking optimization problem defined in one of the above languages I am trying to load the model and it export/convert it to LP solver input format. Model instance creation time is being measured and model instance characteristics are being examined.
So far I have only used sample problems provided by the authors of AMLs which are quite small. In order to have a meaningful benchmark I would like to find test data to build much larger problem model instance.
Maybe someone could guide me where I could find larger data sets for the models of linear programming problems defined in the AMLs mentioned above?