permit Dynare to compute k order perturbation

The current restriction is order ≤ 3.

assigned to @sebastien

mentioned in commit preprocessor@e89dd9cb

Other related commits:

• in the preprocessor: preprocessor@6fa115ae, preprocessor@e3a3992c, preprocessor@571b5d08
• in the k-order DLL: 86a607a4, 2dc988b6

And also many other simplification and modernization commits in Dynare++ and the k-order DLL

mentioned in commit preprocessor@bbab112d

mentioned in commit 4c6e911d

mentioned in commit d7dd7214

mentioned in commit c31ba8be

changed the description

mentioned in commit c10d8daf

mentioned in commit a101457c

mentioned in commit b1ba53ce

mentioned in commit 68aa1ace

mentioned in commit e281f352

Essentially it should now be possible to use an arbitrary approximation order in stoch_simul (up to computational limits).

Feedback welcome.

@JohannesPfeifer It would be great to adapt decision_rules/third_order/FV2011.mod for a higher order. Do you know at which order the Mathematica-computed policy functions are available?

I would need to check. But nowadays I would probably rather cross-check the results with RISE.

I have added in f1b16ced a test of policy functions at order 9 against the Burnside (1998, JEDC) model. It would be good to have a test against a more complex model, but for the time being I’m happy with it.

mentioned in commit 38b588d7

mentioned in commit be170f13

mentioned in commit 1e92e308

mentioned in commit 040c4f11

mentioned in commit preprocessor@c628f212

mentioned in commit f1b16ced

mentioned in commit b2957127

This is essentially done, hence closing.

k-order still needs to be implemented in particle filter, but let's track it elsewhere.

closed