Fixed a bug with running a local method on tasks with partially integ…

…er parameters (#162)

iOpt/method/local_optimizer.py

@@ -36,7 +36,7 @@ def evaluate_function(self, y: List[float]) -> float:
         if self.max_calcs != -1 and self.calcs_count >= self.max_calcs:
             function_value.value = sys.float_info.max
             return function_value.value
-        for i in range(self.task.problem.dimension):
+        for i in range(self.task.problem.number_of_float_variables):
             if (y[i] < self.task.problem.lower_bound_of_float_variables[i]) \
                     or (y[i] > self.task.problem.upper_bound_of_float_variables[i]):
                 function_value.value = sys.float_info.max

test/test_solving_test_problems.py

@@ -1,8 +1,10 @@
 import unittest
 import numpy as np
 from iOpt.problem import Problem
+from iOpt.trial import Point
 from problems.GKLS import GKLS
 from problems.rastrigin import Rastrigin
+from problems.rastriginInt import RastriginInt
 from problems.xsquared import XSquared
 from problems.hill import Hill
 from problems.shekel import Shekel
@@ -15,6 +17,7 @@
 from iOpt.solver_parametrs import SolverParameters
 
 
+
 class TestSolvingProblems(unittest.TestCase):
     """setUp method is overridden from the parent class Rastrigin"""
 
@@ -29,7 +32,7 @@ def checkIsSolved(self, problem: Problem, params: SolverParameters, number_of_gl
         sol = solver.solve()
 
         # Проверяем что найденный АГП минимумом соответствуйте априори известному, для этой задачи, с точностью eps
-        for j in range(problem.dimension):
+        for j in range(problem.number_of_float_variables):
             fabsx = np.abs(problem.known_optimum[0].point.float_variables[j] -
                            sol.best_trials[0].point.float_variables[j])
             fm = params.eps * (problem.upper_bound_of_float_variables[j] -
@@ -38,6 +41,7 @@ def checkIsSolved(self, problem: Problem, params: SolverParameters, number_of_gl
 
         # Проверяем что на решение потребовалось правильное число итераций АГП
         self.assertEqual(sol.number_of_global_trials, number_of_global_trials)
+        return sol
 
     def test_Rastrigin_Solve(self):
         r = 3.5
@@ -47,6 +51,17 @@ def test_Rastrigin_Solve(self):
 
         self.checkIsSolved(problem, params, number_of_global_trials)
 
+    def test_RastriginInt_Solve(self):
+        r = 3.5
+        problem = RastriginInt(dimension=5, number_of_discrete_variables=3)
+        start_point: Point = Point(float_variables=[0.5, 0.5], discrete_variables=['A', 'B', 'A'])
+        params = SolverParameters(r=r, eps=self.epsVal, iters_limit=100000, start_point=start_point,
+                                  number_of_parallel_points=1, refine_solution=True)
+        number_of_global_trials = 5678
+
+        sol = self.checkIsSolved(problem, params, number_of_global_trials)
+        self.assertEqual(sol.number_of_local_trials, 70)
+
     def test_XSquared_Solve(self):
         r = 3.5
         problem = XSquared(1)