"""A class to manage optimization models."""
import pyomo.environ as pyo
from pyomo.opt.base import OptSolver
[docs]
class OptimizationModelManager:
def __init__(self, solver: str | OptSolver, model: pyo.AbstractModel | None = None):
"""Class to manage Pyomo optimization models.
Args:
solver: Pyomo OptSolver, or a string that instantiates a Pyomo OptSolver
with this argument.
model: Pyomo abstract model, optional.
"""
if isinstance(solver, str):
solver = pyo.SolverFactory(solver)
[docs]
self.model_instance: pyo.ConcreteModel | None = None
[docs]
self.is_solved: bool = False
[docs]
def set_model(self, model: pyo.AbstractModel) -> None:
"""Set a new model.
Args:
model: Pyomo abstract model to set.
"""
self.model = model
self.model_instance = None
self.is_solved = False
[docs]
def instantiate(self, data: dict) -> None:
"""Create concrete model with instantiation data.
Args:
data: Dictionary with keys of model attributes and values
of data to be set. The data needs to correspond to the abstract
model in terms of attributes needed and their lengths, otherwise
instantiation will fail.
"""
if self.model is None:
raise RuntimeError("There is no abstract model to instantiate.")
data = {None: data}
self.model_instance = self.model.create_instance(data=data)
self.is_solved = False
[docs]
def solve(self) -> None:
"""Solve the model."""
if not self.model_instance:
raise RuntimeError("No concrete model to solve.")
if self.is_solved:
raise RuntimeError("Model is already solved.")
self.solver.solve(self.model_instance)
self.is_solved = True