This is a small library to simplify solving dense optimization problems with Gurobi in Python. This has been designed to interface with Gurobi such that there is minimal overhead with interacting with the solver. This is not ment for solving large sparse optimization problems, but to solve many small optimization problems quickley.
Online Documentation can be found at Read the Docs!
SimpleDenseGurobi considers problems of the following form.
This interface takes numpy arrays for the problem definitions, and iterables for equality constraint indices and binary variable indices. Any input parameter is nullable. Matrices c, and b should be column vectors.
sol = solve_lp(c, A, b) #solve LP
sol = solve_lp(None, A, b) #testing feasibility
sol = solve_qp(Q, c, A, b) #solve QP
sol = solve_qp(Q, c, None, None) #unconstrained QP
sol = solve_miqp(Q, c, A, b, equality_constraints, binary_variables) #Constrained MIQP with equality constraintsThe solve_XXX interface returns a SolutionOutput object if optimal else None. The SolutionOuput object is generally defined in the following manner.
@dataclass
class SolverOutput:
obj: float
sol: numpy.ndarray
slack: Optional[numpy.ndarray]
active_set: Optional[numpy.ndarray]
dual: Optional[numpy.ndarray]Support for MIQCQP and QCQP problem types are forthcoming.