albop
commented
4 years ago It's not a stupid question. Here is what is happening: decision rules are all defined on a a grid, which is the product of the exogenous and of the endogenous grid. A point in the endogenous one is always referred to by s. If the exogenous grid can be indexed by a discrete array points are indexed by i, or if they are indexed by a continuum of points it is m. Now you're decision rule can be indexed by eval_s (if there is no exogenous grid, or if it's empty), by eval_is or by eval_ms(m,s), when these expressions make sense.
For cubic interpolation, we ignore the first index, if the first grid is empty (that is : eval_is(i,s)=eval_s(s) and eval_ms(m,s)=eval_s(s)).
There is no reason, we wouldn't do the same for linear interpolation. I"ll fix that very soon.
TimMunday
Hi- is there a reason that the dispatch multimethods in decision_rule.py do not include some functions for Linear interpolation with an empty exo_grid. ? Apologies if this is a stupid question.
Under
# Empty x Cartesian x Linear, there are some functions defined but not eval_ms or eval_is, which seem to be needed for impulse response analysis. (And indeed are defined beneath for Empty x Cartesian x Cubic).As a result if one does try to find irfs for a model that has used time iteration and linear interpolation, the following error message occurs:
Could not find signature for eval_ms: <DecisionRule, EmptyGrid, UniformCartesianGrid, Linear, ndarray, ndarray>