Problem classes

There exist specialized solvers for almost every problem class. That is why often three of them end up running side by side in a single project.

Quicopt's API covers all problem classes through a single interface. You find concrete examples with runnable code to get started below.

Linear Programming

Continuous variables, a linear objective, linear constraints.

Continuous variables, a quadratic objective, linear constraints.

A linear model with some integer or binary decisions.

Degree-3+ polynomials, non-smooth penalties, dense or implicit Hessians.

Quadratic unconstrained binary optimization: the native quantum format.

Integer decisions together with a non-linear objective or constraints.

Cost from a simulator, digital twin, or ERP: no formula, no gradients.

Higher-order unconstrained binary optimization: degree 3 and beyond.