Merge pull request #82 from nextmv-io/feature/pyomo-apps-support-scip

Update pyomo apps to support SCIP

python-pyomo-knapsack/main.py

@@ -7,6 +7,7 @@
 SUPPORTED_PROVIDER_DURATIONS = {
     "cbc": "sec",
     "glpk": "tmlim",
+    "scip": "limits/time",
 }
 
 
@@ -84,7 +85,7 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
     model.objective = pyo.Objective(expr=values, sense=pyo.maximize)
 
     # Solves the problem.
-    results = solver.solve(model)
+    results = solver.solve(model, tee=True)
 
     # Convert to solution format.
     value = pyo.value(model.objective, exception=False)

python-pyomo-knapsack/requirements.txt

@@ -1,2 +1,2 @@
-pyomo==6.7.2
+pyomo==6.8.0
 nextmv==0.12.0

python-pyomo-shiftassignment/main.py

@@ -1,5 +1,4 @@
 import datetime
-import logging
 import time
 from typing import Any
 
@@ -10,6 +9,7 @@
 SUPPORTED_PROVIDER_DURATIONS = {
     "cbc": "sec",
     "glpk": "tmlim",
+    "scip": "limits/time",
 }
 
 # Status of the solver after optimizing.
@@ -56,9 +56,6 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
             f"{', '.join(SUPPORTED_PROVIDER_DURATIONS.keys())}"
         )
 
-    # Silence all Pyomo logging.
-    logging.getLogger("pyomo.core").setLevel(logging.ERROR)
-
     # Prepare data
     workers, shifts, rules_per_worker = convert_input(input.data)
 
@@ -146,11 +143,11 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
     )
 
     # Creates the solver.
-    solver = pyo.SolverFactory(provider)  # Use an appropriate solver name
+    solver = pyo.SolverFactory(provider)
     solver.options[SUPPORTED_PROVIDER_DURATIONS[provider]] = options.duration
 
     # Solve the model.
-    results = solver.solve(model)
+    results = solver.solve(model, tee=True)
 
     # Convert to solution format.
     schedule = {}

python-pyomo-shiftassignment/requirements.txt

@@ -1,2 +1,2 @@
-pyomo==6.7.2
+pyomo==6.8.0
 nextmv==0.12.0

python-pyomo-shiftplanning/main.py

@@ -3,24 +3,21 @@
 from typing import Any
 
 import nextmv
-from pyomo.environ import (
-    ConcreteModel,
-    Constraint,
-    NonNegativeIntegers,
-    Objective,
-    SolverFactory,
-    TerminationCondition,
-    Var,
-    minimize,
-    value,
-)
+import pyomo.environ as pyo
+
+# Duration parameter for the solver.
+SUPPORTED_PROVIDER_DURATIONS = {
+    "cbc": "sec",
+    "glpk": "tmlim",
+    "scip": "limits/time",
+}
 
 # Status of the solver after optimizing.
 STATUS = {
-    TerminationCondition.feasible: "suboptimal",
-    TerminationCondition.infeasible: "infeasible",
-    TerminationCondition.optimal: "optimal",
-    TerminationCondition.unbounded: "unbounded",
+    pyo.TerminationCondition.feasible: "suboptimal",
+    pyo.TerminationCondition.infeasible: "infeasible",
+    pyo.TerminationCondition.optimal: "optimal",
+    pyo.TerminationCondition.unbounded: "unbounded",
 }
 
 
@@ -50,8 +47,16 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
     start_time = time.time()
     nextmv.redirect_stdout()  # Solver chatter is logged to stderr.
 
+    # Make sure the provider is supported.
+    provider = options.provider
+    if provider not in SUPPORTED_PROVIDER_DURATIONS:
+        raise ValueError(
+            f"Unsupported provider: {provider}. The supported providers are: "
+            f"{', '.join(SUPPORTED_PROVIDER_DURATIONS.keys())}"
+        )
+
     # Create the Pyomo model
-    model = ConcreteModel()
+    model = pyo.ConcreteModel()
 
     # Prepare data
     shifts, demands = convert_data(input.data)
@@ -67,7 +72,7 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
     required_hours = sum((p.end_time - p.start_time).seconds for p in periods) / 3600
 
     # Create integer variables indicating how many times a shift is planned.
-    model.x_assign = Var([(s["id"],) for s in concrete_shifts], within=NonNegativeIntegers)
+    model.x_assign = pyo.Var([(s["id"],) for s in concrete_shifts], within=pyo.NonNegativeIntegers)
 
     # Bound assignment variables by the minimum and maximum number of workers.
     for s in concrete_shifts:
@@ -77,11 +82,11 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
 
     # Create variables for tracking various costs.
     if "under_supply_cost" in input_options:
-        model.x_under = Var([(p,) for p in periods], within=NonNegativeIntegers)
-        model.underSupply = Var(within=NonNegativeIntegers)
+        model.x_under = pyo.Var([(p,) for p in periods], within=pyo.NonNegativeIntegers)
+        model.underSupply = pyo.Var(within=pyo.NonNegativeIntegers)
     if "over_supply_cost" in input_options:
-        model.overSupply = Var(within=NonNegativeIntegers)
-    model.shift_cost = Var(within=NonNegativeIntegers)
+        model.overSupply = pyo.Var(within=pyo.NonNegativeIntegers)
+    model.shift_cost = pyo.Var(within=pyo.NonNegativeIntegers)
 
     # Objective function: minimize the cost of the planned shifts
     obj_expr = 0
@@ -90,7 +95,7 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
     if "over_supply_cost" in input_options:
         obj_expr += model.overSupply * input_options["over_supply_cost"]
     obj_expr += model.shift_cost
-    model.objective = Objective(expr=obj_expr, sense=minimize)
+    model.objective = pyo.Objective(expr=obj_expr, sense=pyo.minimize)
 
     # Constraints
 
@@ -100,37 +105,40 @@ def solve(input: nextmv.Input, options: nextmv.Options) -> nextmv.Output:
         # Add the new constraint
         model.add_component(
             constraint_name,
-            Constraint(
+            pyo.Constraint(
                 expr=sum([model.x_assign[s["id"]] for s in p.covering_shifts]) == sum(d["count"] for d in p.demands)
             ),
         )
 
     # Track under supply
     if "under_supply_cost" in input_options:
-        model.under_supply = Constraint(
+        model.under_supply = pyo.Constraint(
             expr=model.underSupply
             == sum(model.x_under[p] * (p.end_time - p.start_time).seconds / 3600 for p in periods)
         )
 
     # Track over supply
     if "over_supply_cost" in input_options:
-        model.over_supply = Constraint(
+        model.over_supply = pyo.Constraint(
             expr=model.overSupply
             == sum(model.x_assign[s["id"]] * (s["end_time"] - s["start_time"]).seconds / 3600 for s in concrete_shifts)
             - required_hours
         )
 
     # Track shift cost
-    model.shift_cost_track = Constraint(
+    model.shift_cost_track = pyo.Constraint(
         expr=model.shift_cost == sum(model.x_assign[s["id"]] * s["cost"] for s in concrete_shifts)
     )
 
+    # Creates the solver.
+    solver = pyo.SolverFactory(provider)
+    solver.options[SUPPORTED_PROVIDER_DURATIONS[provider]] = options.duration
+
     # Solve the model.
-    solver = SolverFactory(options.provider)
-    results = solver.solve(model, tee=False, timelimit=options.duration)
+    results = solver.solve(model, tee=True)
 
     # Convert to solution format.
-    val = value(model.objective, exception=False)
+    val = pyo.value(model.objective, exception=False)
     schedule = {
         "planned_shifts": [
             {

python-pyomo-shiftplanning/requirements.txt

@@ -1,2 +1,2 @@
-pyomo==6.7.2
+pyomo==6.8.0
 nextmv==0.12.0