Benoît Legat
Benoît Legat
Some conic solvers might not behave so well with zero rows on the SOC constraints, as this constraint is indeed private to the bridge, the bridge is responsible to do...
> Isn't the alternative the solver throwing an error if it deletes a variable that ends up creating a zero row, and the solver doesn't support zero rows? Why does...
> The solver would throw `DeleteNotAllowed` and the cache would then rebuild. User wouldn't see anything. Conic solvers don't support deletion (except Mosek), it's done in their cache. It's pretty...
Suppose you solve a QP, the problem is long to solve because you have a lot of variables. You delete most of them, you solve again, it still long to...
ECOS doesn't support variable deletion so that won't be an issue with ECOS
It's not about likelihood, a bug is a bug :) We don't know about all solvers or all bridges that exists. If a bug can occur with the current design...
Hi @pekpuglia Good question. The issue here is that we expand everything into an expression graph at the level of elementary operations: `*`, `+`, ... But that graph is way...
Can you post the expression graph you obtain with CasADi (so `F`) after one or two steps ?
Thanks, it seems Casadi is automatically creating the intermediate variable as @odow mentioned. Actually, building the JuMP functions is quite fast. The fact that they have common sub-expressions is not...
Yes, it has nontrivial consequences. It may very well jeopardize our chances of adding a modification API for nonlinear constraints in the future.