fix(genetic): limit EV charging to unmet min SoC

src/akkudoktoreos/optimization/genetic/genetic.py

@@ -983,10 +983,7 @@ class GeneticOptimization(OptimizationBase):
                     "Penalty function parameter `ev_soc_miss` not configured, using {}.", penalty
                 )
             ev_soc_percentage = self.simulation.ev.current_soc_percentage()
-            if (
-                ev_soc_percentage < parameters.eauto.min_soc_percentage
-                or ev_soc_percentage > parameters.eauto.max_soc_percentage
-            ):
+            if ev_soc_percentage < parameters.eauto.min_soc_percentage:
                 gesamtbilanz += (
                     abs(parameters.eauto.min_soc_percentage - ev_soc_percentage) * penalty
                 )
@@ -1020,10 +1017,25 @@ class GeneticOptimization(OptimizationBase):
 
         logger.debug("Start optimize: {}", start_solution)
 
-        # Insert the start solution into the population if provided
+        # Insert the start solution into the population if provided and compatible with the
+        # currently active genome layout. EV optimization adds one gene per prediction hour,
+        # so a cached solution from a previous run without EV optimization must not be reused.
         if start_solution is not None:
-            for _ in range(10):
-                population.insert(0, creator.Individual(start_solution))
+            expected_length = self.config.prediction.hours
+            if self.optimize_ev:
+                expected_length += self.config.prediction.hours
+            if self.opti_param.get("home_appliance", 0) > 0:
+                expected_length += 1
+
+            if len(start_solution) == expected_length:
+                for _ in range(10):
+                    population.insert(0, creator.Individual(start_solution))
+            else:
+                logger.warning(
+                    "Ignoring start_solution with incompatible length {} (expected {}).",
+                    len(start_solution),
+                    expected_length,
+                )
 
         # Run the evolutionary algorithm
         pop, log = algorithms.eaMuPlusLambda(
@@ -1105,7 +1117,7 @@ class GeneticOptimization(OptimizationBase):
             )
             eauto.set_charge_per_hour(np.full(self.config.prediction.hours, 1))
             self.optimize_ev = (
-                parameters.eauto.min_soc_percentage - parameters.eauto.initial_soc_percentage >= 0
+                parameters.eauto.min_soc_percentage > parameters.eauto.initial_soc_percentage
             )
             # electrical vehicle charge rates
             if parameters.eauto.charge_rates is not None:
@@ -1208,11 +1220,9 @@ class GeneticOptimization(OptimizationBase):
         if self.simulation.home_appliance:
             washingstart_int = self.simulation.home_appliance_start_hour
 
-        eautocharge_hours_float = (
-            [self.ev_possible_charge_values[i] for i in eautocharge_hours_index]
-            if eautocharge_hours_index is not None
-            else None
-        )
+        eautocharge_hours_float = None
+        if eautocharge_hours_index is not None and self.simulation.ev is not None:
+            eautocharge_hours_float = self.simulation.ev.charge_array.tolist()
 
         # Simulation may have changed something, use simulation values
         ac_charge_hours = self.simulation.ac_charge_hours

tests/test_inverter_efficiency.py

@@ -591,6 +591,73 @@ def _run_evaluate_with_mocked_sim(
     return fitness[0]
 
 
+def _run_evaluate_with_mocked_ev_soc(config_eos, ev_soc_percentage: float) -> float:
+    """Return fitness for a mocked EV SoC while EV optimization is active."""
+    from akkudoktoreos.optimization.genetic.genetic import GeneticOptimization
+
+    config_eos.merge_settings_from_dict(
+        {
+            "prediction": {"hours": 48},
+            "optimization": {"hours": 48},
+        }
+    )
+    config_eos.optimization.genetic.penalties = {
+        "ev_soc_miss": 10,
+        "ac_charge_break_even": 1.0,
+    }
+
+    optim = GeneticOptimization.__new__(GeneticOptimization)
+    optim.config = config_eos
+    optim.optimize_ev = True
+    optim.verbose = False
+    optim.opti_param = {"home_appliance": 0}
+
+    mock_ev = Mock()
+    mock_ev.current_soc_percentage.return_value = ev_soc_percentage
+
+    sim = Mock()
+    sim.battery = None
+    sim.inverter = None
+    sim.ev = mock_ev
+    sim.ac_charge_hours = None
+    sim.elect_price_hourly = None
+    sim.load_energy_array = None
+    optim.simulation = sim
+
+    dummy_result = {
+        "Gesamtbilanz_Euro": 1.0,
+        "Gesamt_Verluste": 0.0,
+        "EAuto_SoC_pro_Stunde": np.zeros(48),
+    }
+
+    class _Ind(list):  # noqa: N801
+        pass
+
+    fake_individual = _Ind([0] * 96)
+
+    with patch.object(optim, "evaluate_inner", return_value=dummy_result):
+        fitness = optim.evaluate(
+            fake_individual,
+            parameters=Mock(
+                ems=Mock(preis_euro_pro_wh_akku=0.0),
+                eauto=Mock(min_soc_percentage=70, max_soc_percentage=100),
+            ),
+            start_hour=0,
+            worst_case=False,
+        )
+    return fitness[0]
+
+
+class TestEvSocPenalty:
+    """EV SoC target is a minimum-only penalty."""
+
+    def test_ev_soc_below_min_adds_penalty(self, config_eos):
+        assert _run_evaluate_with_mocked_ev_soc(config_eos, 65) == pytest.approx(51.0)
+
+    def test_ev_soc_above_max_does_not_add_min_soc_penalty(self, config_eos):
+        assert _run_evaluate_with_mocked_ev_soc(config_eos, 105) == pytest.approx(1.0)
+
+
 class TestAcChargeBreakEvenPenalty:
     """Break-even penalty in GeneticOptimization.evaluate().