Merge branch 'main' of https://github.com/aimclub/iOpt into main

examples/GKLS_example.py

@@ -30,7 +30,7 @@ def SolveSingleGKLS():
     solver.add_listener(spl)
 
     # Решение задачи
-    sol = solver.solve()
+    solver.solve()
 
 
 if __name__ == "__main__":

examples/GKLS_timeout_example.py

@@ -24,7 +24,7 @@ def SolveSingleGKLS():
     solver.add_listener(cfol)
 
     # Решение задачи
-    sol = solver.solve()
+    solver.solve()
 
 
 if __name__ == "__main__":

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

problems/Floudas.py

@@ -0,0 +1,81 @@
+import numpy as np
+from iOpt.trial import Point
+from iOpt.trial import FunctionValue
+from iOpt.trial import FunctionType
+from iOpt.trial import Trial
+from iOpt.problem import Problem
+import math
+
+
+class Floudas(Problem):
+    """
+
+    """
+
+    def __init__(self):
+        """
+        Конструктор класса Floudas problem.
+        """
+        super(Floudas, self).__init__()
+        self.name = "Floudas"
+        self.dimension = 3
+        self.number_of_float_variables = 2
+        self.number_of_discrete_variables = 1
+        self.number_of_objectives = 1
+        self.number_of_constraints = 3
+
+        self.float_variable_names = np.ndarray(shape=(self.number_of_float_variables,), dtype=object)
+        for i in range(self.number_of_float_variables):
+            self.float_variable_names[i] = str(i)
+
+        self.discrete_variable_names = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        for i in range(self.number_of_discrete_variables):
+            self.discrete_variable_names[i] = str(i)
+
+        self.lower_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.lower_bound_of_float_variables[0] = 0.2
+        self.lower_bound_of_float_variables[1] = -2.22554
+        self.upper_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.upper_bound_of_float_variables[0] = 1
+        self.upper_bound_of_float_variables[1] = -1
+
+        self.discrete_variable_values = [["0", "1"] for i in range(self.number_of_discrete_variables)]
+
+        self.known_optimum = np.ndarray(shape=(1,), dtype=Trial)
+
+        pointfv = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        pointfv = [0.499609, -1.305787]
+
+        pointdv = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        pointdv[0] = "1"
+
+        KOpoint = Point(pointfv, pointdv)
+        KOfunV = np.ndarray(shape=(1,), dtype=FunctionValue)
+        KOfunV[0] = FunctionValue()
+        KOfunV[0].value = 0.100001
+        self.known_optimum[0] = Trial(KOpoint, KOfunV)
+
+
+    def calculate(self, point: Point, function_value: FunctionValue) -> FunctionValue:
+        """
+        Вычисление значения выбранной функции в заданной точке.
+
+        :param point: координаты точки испытания, в которой будет вычислено значение функции
+        :param function_value: объект определяющий номер функции в задаче и хранящий значение функции
+        :return: Вычисленное значение функции в точке point
+        """
+        result: np.double = 0
+        x = point.float_variables
+        b = point.discrete_variables
+
+        if function_value.type == FunctionType.OBJECTIV:
+            result = np.double(-0.7 * int(b[0]) + 5.0 * (x[0] - 0.5) * (x[0] - 0.5) + 0.8)
+        elif function_value.functionID == 0:  # constraint 1
+            result = np.double(-math.exp(x[0] - 0.2) - x[1])
+        elif function_value.functionID == 1:  # constraint 2
+            result = np.double(x[1] + 1.1 * int(b[0]) - 1)
+        elif function_value.functionID == 2:  # constraint 3
+            result = np.double(x[0] - 1.2 * int(b[0]) - 0.2)
+
+        function_value.value = result
+        return function_value

problems/GKLS_function/gkls_function.py

@@ -1,5 +1,4 @@
 import numpy as np
-import math
 from problems.GKLS_function.gkls_random import GKLSRandomGenerator
 
 

problems/Pern.py

@@ -0,0 +1,80 @@
+import numpy as np
+from iOpt.trial import Point
+from iOpt.trial import FunctionValue
+from iOpt.trial import FunctionType
+from iOpt.trial import Trial
+from iOpt.problem import Problem
+import math
+
+
+class Pern(Problem):
+    """
+
+    """
+
+    def __init__(self):
+        """
+        Конструктор класса Pern problem.
+        """
+        super(Pern, self).__init__()
+        self.name = "Pern"
+        self.dimension = 2
+        self.number_of_float_variables = 1
+        self.number_of_discrete_variables = 1
+        self.number_of_objectives = 1
+        self.number_of_constraints = 3
+
+        self.float_variable_names = np.ndarray(shape=(self.number_of_float_variables,), dtype=object)
+        for i in range(self.number_of_float_variables):
+            self.float_variable_names[i] = str(i)
+
+        self.discrete_variable_names = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        for i in range(self.number_of_discrete_variables):
+            self.discrete_variable_names[i] = str(i)
+
+        self.lower_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.lower_bound_of_float_variables[0] = 1
+        self.upper_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.upper_bound_of_float_variables[0] = 10
+
+        self.discrete_variable_values = [[[str(i) for i in range(1, 7)]] for i in range(self.number_of_discrete_variables)]
+
+        self.known_optimum = np.ndarray(shape=(1,), dtype=Trial)
+
+        pointfv = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        pointfv[0] = 4
+
+        pointdv = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        pointdv[0] = "1"
+
+        KOpoint = Point(pointfv, pointdv)
+        KOfunV = np.ndarray(shape=(1,), dtype=FunctionValue)
+        KOfunV[0] = FunctionValue()
+        KOfunV[0].value = -17
+        self.known_optimum[0] = Trial(KOpoint, KOfunV)
+
+
+    def calculate(self, point: Point, function_value: FunctionValue) -> FunctionValue:
+        """
+        Вычисление значения выбранной функции в заданной точке.
+
+        :param point: координаты точки испытания, в которой будет вычислено значение функции
+        :param function_value: объект определяющий номер функции в задаче и хранящий значение функции
+        :return: Вычисленное значение функции в точке point
+        """
+        result: np.double = 0
+        x = point.float_variables[0]
+        b = int(point.discrete_variables[0])
+
+        if function_value.type == FunctionType.OBJECTIV:
+            result = np.double(3.0 * b - 5.0 * x)
+        elif function_value.functionID == 0:  # constraint 1
+            result = np.double(2.0 * b * b - 2.0 * math.sqrt(b) - 2.0 * math.sqrt(x) * b * b
+                               + 11.0 * b + 8 * x - 39.0)
+        elif function_value.functionID == 1:  # constraint 2
+            result = np.double(-b + x - 3.0)
+        elif function_value.functionID == 2:  # constraint 3
+            result = np.double(2.0 * b + 3 * x - 24.0)
+
+        function_value.value = result
+        return function_value

problems/Synthes.py

@@ -0,0 +1,95 @@
+import numpy as np
+from iOpt.trial import Point
+from iOpt.trial import FunctionValue
+from iOpt.trial import FunctionType
+from iOpt.trial import Trial
+from iOpt.problem import Problem
+import math
+
+
+class Synthes(Problem):
+    """
+
+    """
+
+    def __init__(self):
+        """
+        Конструктор класса Synthes problem.
+        """
+        super(Synthes, self).__init__()
+        self.name = "Synthes"
+        self.dimension = 6
+        self.number_of_float_variables = 3
+        self.number_of_discrete_variables = 3
+        self.number_of_objectives = 1
+        self.number_of_constraints = 6
+
+        self.float_variable_names = np.ndarray(shape=(self.number_of_float_variables,), dtype=object)
+        for i in range(self.number_of_float_variables):
+            self.float_variable_names[i] = str(i)
+
+        self.discrete_variable_names = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        for i in range(self.number_of_discrete_variables):
+            self.discrete_variable_names[i] = str(i)
+
+        self.lower_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.lower_bound_of_float_variables.fill(0)
+        self.upper_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.upper_bound_of_float_variables.fill(3)
+
+        self.discrete_variable_values = [[[str(i) for i in range(0, 6)]] for i in range(self.number_of_discrete_variables)]
+
+        self.known_optimum = np.ndarray(shape=(1,), dtype=Trial)
+        # UNDEFINED
+
+
+    def calculate(self, point: Point, function_value: FunctionValue) -> FunctionValue:
+        """
+        Вычисление значения выбранной функции в заданной точке.
+
+        :param point: координаты точки испытания, в которой будет вычислено значение функции
+        :param function_value: объект определяющий номер функции в задаче и хранящий значение функции
+        :return: Вычисленное значение функции в точке point
+        """
+        result: np.double = 0
+        x = point.float_variables
+        b = [int(x) for x in point.discrete_variables]
+
+        if function_value.type == FunctionType.OBJECTIV:
+            result = np.double(5.0 * b[0] + 6.0 * b[1] + 8.0 * b[2] + 10.0 * x[0]
+                               - 7.0 * x[2] - 18.0 * math.log(x[1] + 1.0)
+                               - 19.2 * math.log(x[0] - x[1] + 1.0) + 10.0)
+        elif function_value.functionID == 0:  # constraint 1
+            result = np.double(b[0] + b[1] - 1.1)
+        elif function_value.functionID == 1:  # constraint 2
+            if ((x[0] - x[1] + 1.0) != 0):
+                try:
+                    result = np.double(-(math.log(x[1] + 1.0) + 1.2*
+                                     math.log(x[0] - x[1] + 1.0) - x[2]- 2 * b[2] + 2.0))
+                except ValueError:
+                    print("CalculateFuncs Error!!!")
+                    result = np.NaN
+                    pass  # do nothing!
+            else:
+                result = 1
+        elif function_value.functionID == 2:  # constraint 3
+            if ((x[0] - x[1] + 1.0) != 0):
+                try:
+                    result = np.double(-(math.log(x[1] + 1.0) + 1.2*
+                                     math.log(x[0] - x[1] + 1.0) - x[2]- 2 * b[2] + 2.0))
+                except ValueError:
+                    print("CalculateFuncs Error!!!")
+                    result = np.NaN
+                    pass  # do nothing!
+
+            else:
+                result = 1
+        elif function_value.functionID == 3:  # constraint 4
+            result = np.double(x[1] - x[0])
+        elif function_value.functionID == 4:  # constraint 5
+            result = np.double(x[1] - 2.0 * b[0])
+        elif function_value.functionID == 5:  # constraint 6
+            result = np.double(x[0] - x[1] - 2.0 * b[1])
+
+        function_value.value = result
+        return function_value

problems/Yuan.py

@@ -0,0 +1,94 @@
+import numpy as np
+from iOpt.trial import Point
+from iOpt.trial import FunctionValue
+from iOpt.trial import FunctionType
+from iOpt.trial import Trial
+from iOpt.problem import Problem
+import math
+
+
+class Yuan(Problem):
+    """
+
+    """
+
+    def __init__(self):
+        """
+        Конструктор класса Yuan problem.
+        """
+        super(Yuan, self).__init__()
+        self.name = "Yuan"
+        self.dimension = 7
+        self.number_of_float_variables = 3
+        self.number_of_discrete_variables = 4
+        self.number_of_objectives = 1
+        self.number_of_constraints = 9
+
+        self.float_variable_names = np.ndarray(shape=(self.number_of_float_variables,), dtype=object)
+        for i in range(self.number_of_float_variables):
+            self.float_variable_names[i] = str(i)
+
+        self.discrete_variable_names = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        for i in range(self.number_of_discrete_variables):
+            self.discrete_variable_names[i] = str(i)
+
+        self.lower_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.lower_bound_of_float_variables.fill(0)
+        self.upper_bound_of_float_variables = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        self.upper_bound_of_float_variables.fill(3)
+
+        self.discrete_variable_values = [["0", "1"] for i in range(self.number_of_discrete_variables)]
+
+        self.known_optimum = np.ndarray(shape=(1,), dtype=Trial)
+
+        pointfv = np.ndarray(shape=(self.number_of_float_variables,), dtype=np.double)
+        pointfv = [0.2, 0.8, 1.908]
+
+        pointdv = np.ndarray(shape=(self.number_of_discrete_variables,), dtype=object)
+        pointdv = ["1", "1", "0", "1"]
+
+        KOpoint = Point(pointfv, pointdv)
+        KOfunV = np.ndarray(shape=(1,), dtype=FunctionValue)
+        KOfunV[0] = FunctionValue()
+        KOfunV[0].value = 4.5796
+        self.known_optimum[0] = Trial(KOpoint, KOfunV)
+
+
+    def calculate(self, point: Point, function_value: FunctionValue) -> FunctionValue:
+        """
+        Вычисление значения выбранной функции в заданной точке.
+
+        :param point: координаты точки испытания, в которой будет вычислено значение функции
+        :param function_value: объект определяющий номер функции в задаче и хранящий значение функции
+        :return: Вычисленное значение функции в точке point
+        """
+        result: np.double = 0
+        x = point.float_variables
+        b = [int(x) for x in point.discrete_variables]
+
+        if function_value.type == FunctionType.OBJECTIV:
+            result = np.double((b[0] - 1.0) * (b[0] - 1.0) + (b[1] - 2.0) * (b[1] - 2.0) +
+                               (b[2] - 1.0) * (b[2] - 1.0) - math.log(b[3] + 1.0) +
+                               (x[0] - 1.0) * (x[0] - 1.0) + (x[1] - 2.0) * (x[1] - 2.0) +
+                               (x[2] - 3.0) * (x[2] - 3.0))
+        elif function_value.functionID == 0:  # constraint 1
+            result = np.double(b[0] + b[1] + b[2] + x[0] + x[1] + x[2] - 5.0)
+        elif function_value.functionID == 1:  # constraint 2
+            result = np.double(b[2] * b[2] + x[0] * x[0] + x[1] * x[1] + x[2] * x[2] - 5.5)
+        elif function_value.functionID == 2:  # constraint 3
+            result = np.double(b[0] + x[0] - 1.2)
+        elif function_value.functionID == 3:  # constraint 4
+            result = np.double(b[1] + x[1] - 1.8)
+        elif function_value.functionID == 4:  # constraint 5
+            result = np.double(b[2] + x[2] - 2.5)
+        elif function_value.functionID == 5:  # constraint 6
+            result = np.double(b[3] + x[0] - 1.2)
+        elif function_value.functionID == 6:  # constraint 7
+            result = np.double(b[1] * b[1] + x[1] * x[1] - 1.64)
+        elif function_value.functionID == 7:  # constraint 8
+            result = np.double(b[2] * b[2] + x[2] * x[2] - 4.25)
+        elif function_value.functionID == 8:  # constraint 9
+            result = np.double(b[1] * b[1] + x[2] * x[2] - 4.64)
+
+        function_value.value = result
+        return function_value