ludwigkr / optimization-design-library Goto Github PK

Currently I convert the optimization problem of casadi to text and parse the text in the form I want.
This has the disadvantage, that variables can get confused in the parsing process.

This is especially a danger if one variable name is similar to other variable name. E.g. P and SP, in the replacing process while replacing P with the common param word, then also SP is changed to Sparam which is noting anymore.

Currently, the constraint limits don't use the constraint limit correction through the linearization of the constraints.

Currently, all formulation are parsed in sparse matrices.
This has the advantage of a general description for all forms (single expression, vectors, matrices).
However, for large optimization problems, the expression in every element get quite long and repeat themselves over multiple elements.
In these cases it would be better to first calculate some in-between variables which are used multiple time later on.