Ideally, dummy imports should be deactivated in healthy models (by un-checking NRG, DEM etc under Tools - user options, or via START of IMP*Z processes). However, they are often convenient to have around, for various reasons.
My thumb rule for assigning costs to dummy imports is 10x the highest price of a commodity in each category. For example, if the highest shadow price of DEM commodities is 250, then I would use 2500 as the cost of IMPDEMZ. Assigning very high costs can create numerical issues for solvers.
However, IMPNRGZ produces dummy commodities for user constraints as well, and some can have much higher shadow prices due to different units. For example, user constraints that are based on CAP or NCAP.
The current version of Veda allows parameter declarations for dummy commodities that are created for each user constraint (_ and -) via any transformation table. For example, if you wanted to disable the dummy variables for a certain group of user constraints, you could declare FLO_BND(IMPNRGZ,FX) for commodity = ? as 2 for year=0.
My thumb rule for assigning costs to dummy imports is 10x the highest price of a commodity in each category. For example, if the highest shadow price of DEM commodities is 250, then I would use 2500 as the cost of IMPDEMZ. Assigning very high costs can create numerical issues for solvers.
However, IMPNRGZ produces dummy commodities for user constraints as well, and some can have much higher shadow prices due to different units. For example, user constraints that are based on CAP or NCAP.
The current version of Veda allows parameter declarations for dummy commodities that are created for each user constraint (