Require a callable as the internal solver

docs/customization.md

@@ -4,13 +4,16 @@
 
 ### Selecting the internal MIP solver
 
-By default, `LearningSolver` uses [Gurobi](https://www.gurobi.com/) as its internal MIP solver. Another supported solver is [IBM ILOG CPLEX](https://www.ibm.com/products/ilog-cplex-optimization-studio). To switch between solvers, use the `solver` constructor argument, as shown below. It is also possible to specify a time limit (in seconds) and a relative MIP gap tolerance.
+By default, `LearningSolver` uses [Gurobi](https://www.gurobi.com/) as its internal MIP solver, and expects models to be provided using the Pyomo modeling language. Other supported solvers and modeling languages include:
+
+* `CplexPyomoSolver`: [IBM ILOG CPLEX](https://www.ibm.com/products/ilog-cplex-optimization-studio) with Pyomo. 
+* `GurobiSolver`: Gurobi without any modeling language.
+
+To switch between solvers, provide the desired class using the `solver` argument:
 
 ```python
-from miplearn import LearningSolver
+from miplearn import LearningSolver, CplexPyomoSolver
-solver = LearningSolver(solver="cplex",
+solver = LearningSolver(solver=CplexPyomoSolver)
-                        time_limit=300,
-                        gap_tolerance=1e-3)
 ```
 
 ## Customizing solver components

miplearn/components/steps/tests/test_convert_tight.py

@@ -13,7 +13,7 @@ def test_convert_tight_usage():
         capacity=16.0,
     )
     solver = LearningSolver(
-        solver=GurobiSolver(),
+        solver=GurobiSolver,
         components=[
             RelaxIntegralityStep(),
             ConvertTightIneqsIntoEqsStep(),
@@ -64,7 +64,7 @@ def test_convert_tight_infeasibility():
     comp.classifiers["c3"].predict_proba = Mock(return_value=[[1, 0]])
 
     solver = LearningSolver(
-        solver=GurobiSolver(params={}),
+        solver=GurobiSolver,
         components=[comp],
         solve_lp_first=False,
     )
@@ -87,7 +87,7 @@ def test_convert_tight_suboptimality():
     comp.classifiers["c3"].predict_proba = Mock(return_value=[[0, 1]])
 
     solver = LearningSolver(
-        solver=GurobiSolver(params={}),
+        solver=GurobiSolver,
         components=[comp],
         solve_lp_first=False,
     )
@@ -110,7 +110,7 @@ def test_convert_tight_optimal():
     comp.classifiers["c3"].predict_proba = Mock(return_value=[[0, 1]])
 
     solver = LearningSolver(
-        solver=GurobiSolver(params={}),
+        solver=GurobiSolver,
         components=[comp],
         solve_lp_first=False,
     )

miplearn/problems/tests/test_tsp.py

@@ -36,18 +36,17 @@ def test_instance():
         ]
     )
     instance = TravelingSalesmanInstance(n_cities, distances)
-    for solver_name in ["gurobi"]:
+    solver = LearningSolver()
-        solver = LearningSolver(solver=solver_name)
+    solver.solve(instance)
-        solver.solve(instance)
+    x = instance.solution["x"]
-        x = instance.solution["x"]
+    assert x[0, 1] == 1.0
-        assert x[0, 1] == 1.0
+    assert x[0, 2] == 0.0
-        assert x[0, 2] == 0.0
+    assert x[0, 3] == 1.0
-        assert x[0, 3] == 1.0
+    assert x[1, 2] == 1.0
-        assert x[1, 2] == 1.0
+    assert x[1, 3] == 0.0
-        assert x[1, 3] == 0.0
+    assert x[2, 3] == 1.0
-        assert x[2, 3] == 1.0
+    assert instance.lower_bound == 4.0
-        assert instance.lower_bound == 4.0
+    assert instance.upper_bound == 4.0
-        assert instance.upper_bound == 4.0
 
 
 def test_subtour():
@@ -64,17 +63,16 @@ def test_subtour():
     )
     distances = squareform(pdist(cities))
     instance = TravelingSalesmanInstance(n_cities, distances)
-    for solver_name in ["gurobi"]:
+    solver = LearningSolver()
-        solver = LearningSolver(solver=solver_name)
+    solver.solve(instance)
-        solver.solve(instance)
+    assert hasattr(instance, "found_violated_lazy_constraints")
-        assert hasattr(instance, "found_violated_lazy_constraints")
+    assert hasattr(instance, "found_violated_user_cuts")
-        assert hasattr(instance, "found_violated_user_cuts")
+    x = instance.solution["x"]
-        x = instance.solution["x"]
+    assert x[0, 1] == 1.0
-        assert x[0, 1] == 1.0
+    assert x[0, 4] == 1.0
-        assert x[0, 4] == 1.0
+    assert x[1, 2] == 1.0
-        assert x[1, 2] == 1.0
+    assert x[2, 3] == 1.0
-        assert x[2, 3] == 1.0
+    assert x[3, 5] == 1.0
-        assert x[3, 5] == 1.0
+    assert x[4, 5] == 1.0
-        assert x[4, 5] == 1.0
+    solver.fit([instance])
-        solver.fit([instance])
+    solver.solve(instance)
-        solver.solve(instance)

miplearn/solvers/learning.py

@@ -20,8 +20,8 @@ from .. import (
     DynamicLazyConstraintsComponent,
     UserCutsComponent,
 )
-from .pyomo.cplex import CplexPyomoSolver
+from ..solvers.internal import InternalSolver
-from .pyomo.gurobi import GurobiPyomoSolver
+from ..solvers.pyomo.gurobi import GurobiPyomoSolver
 
 logger = logging.getLogger(__name__)
 
@@ -52,33 +52,22 @@ class LearningSolver:
 
     Parameters
     ----------
-    components
+    components: [Component]
         Set of components in the solver. By default, includes:
             - ObjectiveValueComponent
             - PrimalSolutionComponent
             - DynamicLazyConstraintsComponent
             - UserCutsComponent
-    gap_tolerance
+    mode: str
-        Relative MIP gap tolerance. By default, 1e-4.
-    mode
         If "exact", solves problem to optimality, keeping all optimality
         guarantees provided by the MIP solver. If "heuristic", uses machine
         learning more aggressively, and may return suboptimal solutions.
-    solver
+    solver: Callable[[], InternalSolver]
-        The internal MIP solver to use. Can be either "cplex", "gurobi", a
+        A callable that constructs the internal solver. If None is provided,
-        solver class such as GurobiSolver, or a solver instance such as
+        use GurobiPyomoSolver.
-        GurobiSolver().
+    use_lazy_cb: bool
-    threads
+        If true, use native solver callbacks for enforcing lazy constraints,
-        Maximum number of threads to use. If None, uses solver default.
+        instead of a simple loop. May not be supported by all solvers.
-    time_limit
-        Maximum running time in seconds. If None, uses solver default.
-    node_limit
-        Maximum number of branch-and-bound nodes to explore. If None, uses
-        solver default.
-    use_lazy_cb
-        If True, uses lazy callbacks to enforce lazy constraints, instead of
-        a simple solver loop. This functionality may not supported by
-        all internal MIP solvers.
     solve_lp_first: bool
         If true, solve LP relaxation first, then solve original MILP. This
         option should be activated if the LP relaxation is not very
@@ -94,27 +83,23 @@ class LearningSolver:
     def __init__(
         self,
         components=None,
-        gap_tolerance=1e-4,
         mode="exact",
-        solver="gurobi",
+        solver=None,
-        threads=None,
-        time_limit=None,
-        node_limit=None,
-        solve_lp_first=True,
         use_lazy_cb=False,
+        solve_lp_first=True,
         simulate_perfect=False,
     ):
+        if solver is None:
+            solver = GurobiPyomoSolver
+        assert callable(solver), f"Callable expected. Found {solver.__class__} instead."
+
         self.components = {}
         self.mode = mode
         self.internal_solver = None
-        self.internal_solver_factory = solver
+        self.solver_factory = solver
-        self.threads = threads
-        self.time_limit = time_limit
-        self.gap_tolerance = gap_tolerance
-        self.tee = False
-        self.node_limit = node_limit
-        self.solve_lp_first = solve_lp_first
         self.use_lazy_cb = use_lazy_cb
+        self.tee = False
+        self.solve_lp_first = solve_lp_first
         self.simulate_perfect = simulate_perfect
 
         if components is not None:
@@ -130,30 +115,6 @@ class LearningSolver:
         for component in self.components.values():
             component.mode = self.mode
 
-    def _create_internal_solver(self):
-        logger.debug("Initializing %s" % self.internal_solver_factory)
-        if self.internal_solver_factory == "cplex":
-            solver = CplexPyomoSolver()
-        elif self.internal_solver_factory == "gurobi":
-            solver = GurobiPyomoSolver()
-        elif callable(self.internal_solver_factory):
-            solver = self.internal_solver_factory()
-        else:
-            solver = self.internal_solver_factory
-        if self.threads is not None:
-            logger.info("Setting threads to %d" % self.threads)
-            solver.set_threads(self.threads)
-        if self.time_limit is not None:
-            logger.info("Setting time limit to %f" % self.time_limit)
-            solver.set_time_limit(self.time_limit)
-        if self.gap_tolerance is not None:
-            logger.info("Setting gap tolerance to %f" % self.gap_tolerance)
-            solver.set_gap_tolerance(self.gap_tolerance)
-        if self.node_limit is not None:
-            logger.info("Setting node limit to %d" % self.node_limit)
-            solver.set_node_limit(self.node_limit)
-        return solver
-
     def solve(
         self,
         instance,
@@ -255,7 +216,7 @@ class LearningSolver:
                 model = instance.to_model()
 
         self.tee = tee
-        self.internal_solver = self._create_internal_solver()
+        self.internal_solver = self.solver_factory()
         self.internal_solver.set_instance(instance, model)
 
         if self.solve_lp_first:

miplearn/solvers/tests/test_learning_solver.py

@@ -24,9 +24,6 @@ def test_learning_solver():
             logger.info("Solver: %s" % internal_solver)
             instance = _get_instance(internal_solver)
             solver = LearningSolver(
-                time_limit=300,
-                gap_tolerance=1e-3,
-                threads=1,
                 solver=internal_solver,
                 mode=mode,
             )
@@ -115,7 +112,7 @@ def test_solve_fit_from_disk():
 
 
 def test_simulate_perfect():
-    internal_solver = GurobiSolver()
+    internal_solver = GurobiSolver
     instance = _get_instance(internal_solver)
     with tempfile.NamedTemporaryFile(suffix=".pkl", delete=False) as tmp:
         pickle.dump(instance, tmp)
@@ -124,6 +121,5 @@ def test_simulate_perfect():
             solver=internal_solver,
             simulate_perfect=True,
         )
-
         stats = solver.solve(tmp.name)
         assert stats["Lower bound"] == stats["Predicted LB"]