#2 Fixing Github issue 13

ddmpc/controller/model_predictive/nlp.py

@@ -1,8 +1,10 @@
 import time
+import warnings
 from abc import ABC, abstractmethod
 from typing import Optional, Iterator
 from typing import Union
 
+import numpy as np
 import pandas as pd
 from casadi import MX, inf, nlpsol, vertcat
 
@@ -670,6 +672,31 @@ def _add_objectives(self, *objectives: Objective):
 
  self._objectives.append(NLPObjective(objective(nlp_value.mx)))
 
+ def _get_coldstart(self):
+
+ cold_start_values: list = list()
+ for opt_var in self._opt_vars:
+
+ if isinstance(opt_var.feature, Control):
+ cold_start_values.append(opt_var.feature.default)
+
+ elif isinstance(opt_var.feature, Controlled):
+
+ lb, ub = opt_var.feature.mode.bounds(0)
+ target = opt_var.feature.mode.target(0)
+
+ if target is not np.NAN:
+ cold_start_values.append(target)
+ elif lb is not np.NAN and ub is not np.NAN:
+ cold_start_values.append((lb + ub) / 2)
+ else:
+ cold_start_values.append(0)
+ warnings.warn(f'Mode for feature {opt_var.feature.source.name} does not provide lb, ub or target for cold start initialization, using 0 as default.')
+ else:
+ cold_start_values.append(0)
+ warnings.warn(f'Cold start initialization for class {opt_var.feature.source} not implemented yet, using 0 as default.')
+ return np.array(cold_start_values)
+
  def summary(self):
 
  print('------------------------- NLP SUMMARY -------------------------')
@@ -750,6 +777,8 @@ def solve(self, par_vals: list[float]) -> NLPSolution:
  # warm start if a solution is available
  if self.solution is not None and self.lastSolutionFailed is False:
  nlp_instance['x0'] = self.solution.opt_vals
+ else:
+ nlp_instance['x0'] = self._get_coldstart()
 
 
  # call to the solver