Merge pull request #53 from PasaOpasen/7v

v6.7.6

README.md

@@ -14,6 +14,7 @@ version](https://badge.fury.io/py/geneticalgorithm2.svg)](https://pypi.org/proje
 - [Installation](#installation)
 - [Updates information](#updates-information)
   - [**Future**](#future)
+  - [6.7.6 bug fix](#676-bug-fix)
   - [6.7.5 refactor](#675-refactor)
   - [6.7.4 bug fix](#674-bug-fix)
   - [6.7.3 speed up](#673-speed-up)
@@ -130,6 +131,12 @@ pip3 install geneticalgorithm2
 - `function_timeout` and `function` will be moved to `run()` method
 - new stop criteria callbacks (min std, max functions evaluations)
 
+## 6.7.6 bug fix
+
+- fix some bug of `variable_type=='bool'`
+- some refactor of progress bar
+- add some dependences to `setup.py`
+
 ## 6.7.5 refactor
 
 - shorter progress bar (length can be controled by setting `PROGRESS_BAR_LEN` field of `geneticalgorithm2` class)
@@ -1167,7 +1174,7 @@ There is an amazing way to control optimization process using `MiddleCallbacks`
         selection=self.selection,
 
         current_stagnation=counter,
-        max_stagnation=self.mniwi,
+        max_stagnation=self.count_stagnation,
 
         parents_portion=self.param.parents_portion,
         elit_ratio=self.param.elit_ratio,

geneticalgorithm2/geneticalgorithm2.py

@@ -33,7 +33,7 @@
 import sys
 import time
 import random
-
+import math
 
 import numpy as np
 from func_timeout import func_timeout, FunctionTimedOut
@@ -47,7 +47,7 @@
 from .initializer import Population_initializer
 from .plotting_tools import plot_pop_scores, plot_several_lines
 
-from .utils import can_be_prob, is_numpy, union_to_matrix
+from .utils import can_be_prob, is_numpy, union_to_matrix, fast_max
 
 #endregion
 
@@ -236,7 +236,7 @@ def __set_var_boundaries(
         variable_type: Union[str, Sequence[str]],
         variable_boundaries
     ):
-        if variable_type == 'bool':
+        if isinstance(variable_type, str) and variable_type == 'bool':
             self.var_bound = np.array([[0, 1]] * self.dim)
         else:
 
@@ -272,12 +272,11 @@ def __set_max_iterations(self):
         else:
             self.iterate = int(self.param['max_num_iteration'])
 
-        self.stop_mniwi = False  # is stopped cuz of no progress some iterations
         max_it = self.param['max_iteration_without_improv']
         if max_it is None:
-            self.mniwi = self.iterate + 1
+            self.count_stagnation = self.iterate + 1
         else:
-            self.mniwi = int(max_it)
+            self.count_stagnation = int(max_it)
 
     #endregion
 
@@ -489,7 +488,7 @@ def get_data():
                 selection=self.selection,
 
                 current_stagnation=fails,
-                max_stagnation=self.mniwi,
+                max_stagnation=self.count_stagnation,
 
                 parents_portion=self.param.parents_portion,
                 elit_ratio=self.param.elit_ratio,
@@ -524,7 +523,7 @@ def set_data(data: MiddleCallbackData):
 
             fails = data.current_stagnation
             reason_to_stop = data.reason_to_stop
-            self.mniwi = data.max_stagnation
+            self.count_stagnation = data.max_stagnation
 
             self.set_function = data.set_function
 
@@ -758,7 +757,7 @@ def revolution(pop: np.ndarray, scores: np.ndarray, stagnation_count: int) -> Tu
             scores = scores[args_to_sort]
             self._update_report(scores)
 
-            if fails > self.mniwi:
+            if fails > self.count_stagnation:
                 reason_to_stop = f"limit of fails: {fails}"
             elif t == self.iterate:
                 reason_to_stop = f'limit of iterations: {t}'
@@ -771,20 +770,18 @@ def revolution(pop: np.ndarray, scores: np.ndarray, stagnation_count: int) -> Tu
                 show_progress(t, self.iterate, f"GA is running... STOP! {reason_to_stop}")
                 break
 
-            show_progress(t, self.iterate, f"GA is running...{t} gen from {self.iterate}")
-
-
             if scores[0] < self.best_function:  # if there is progress
                 fails = 0
                 self.best_function = scores[0]
-
-                show_progress(t,
-                              self.iterate,
-                              f"GA is running...{t} gen from {self.iterate}...best value = {self.best_function}"
-                              )
             else:
                 fails += 1
 
+            show_progress(
+                t,
+                self.iterate,
+                f"GA is running...{t} gen from {self.iterate}...best value = {self.best_function}"
+            )
+
             # Select parents
 
             par = np.empty((self.par_s, self.dim))
@@ -802,14 +799,16 @@ def revolution(pop: np.ndarray, scores: np.ndarray, stagnation_count: int) -> Tu
             par[par_slice] = pop[new_par_inds].copy()
             par_scores[par_slice] = scores[new_par_inds].copy()
 
+            #
             #  select parents for crossover
-            ef_par_list = np.random.random(self.par_s) < self.prob_cross
-            par_count = ef_par_list.sum()
-            if par_count < 2:
-                while par_count < 2:  # 2 parents at least
-                    ef_par_list = np.random.random(self.par_s) <= self.prob_cross
-                    par_count = np.sum(ef_par_list)
-
+            #
+            par_count = fast_max(2, math.floor(self.par_s * self.prob_cross))  # at least 2 parents
+            #  select random parents
+            ef_par_list = np.random.choice(
+                np.arange(self.par_s),
+                par_count,
+                replace=False
+            )
             ef_par = par[ef_par_list].copy()
 
             # New generation

geneticalgorithm2/mutations.py

@@ -5,7 +5,7 @@
 import numpy as np
 
 
-from .utils import min_fast, max_fast
+from .utils import fast_min, fast_max
 
 
 class Mutations:
@@ -29,7 +29,7 @@ def mutations_discrete_dict():
     def uniform_by_x():
 
         def func(x: float, left: float, right: float):
-            alp = min_fast(x - left, right - x)
+            alp = fast_min(x - left, right - x)
             return random.uniform(x - alp, x + alp)
         return func
 
@@ -48,7 +48,7 @@ def gauss_by_x(sd: float = 0.3):
         """
         def func(x: float, left: float, right: float):
             std = sd * (right - left)
-            return max_fast(left, min_fast(right, np.random.normal(loc = x, scale = std)))
+            return fast_max(left, fast_min(right, np.random.normal(loc = x, scale = std)))
 
         return func
 
@@ -59,7 +59,7 @@ def gauss_by_center(sd: float = 0.3):
         """
         def func(x: float, left: float, right: float):
             std = sd * (right - left)
-            return max_fast(left, min_fast(right, np.random.normal(loc = (left+right)*0.5, scale = std)))
+            return fast_max(left, fast_min(right, np.random.normal(loc =(left + right) * 0.5, scale = std)))
 
         return func
 

geneticalgorithm2/utils.py

@@ -8,13 +8,14 @@
 import numpy as np
 
 
-def min_fast(a, b):
+def fast_min(a, b):
     '''
     1.5 times faster than row min(a, b)
     '''
-    return  a if a < b else b
+    return a if a < b else b
 
-def max_fast(a, b):
+
+def fast_max(a, b):
     return a if a > b else b
 
 
@@ -35,7 +36,7 @@ def split_matrix(mat: np.ndarray):
     """
     splits wide pop matrix to variables and scores
     """
-    return mat[:,:-1], mat[:, -1]
+    return mat[:, :-1], mat[:, -1]
 
 
 def union_to_matrix(variables_2D: np.ndarray, scores_1D: np.ndarray):

setup.py

@@ -6,14 +6,15 @@
 
 setuptools.setup(
     name="geneticalgorithm2", 
-    version="6.7.5",
+    version="6.7.6",
     author="Demetry Pascal",
     author_email="qtckpuhdsa@gmail.com",
     maintainer='Demetry Pascal',
     description="Supported highly optimized and flexible genetic algorithm package for python",
     long_description=long_description,
     long_description_content_type="text/markdown",
     url="https://github.com/PasaOpasen/geneticalgorithm2",
+    license='MIT',
     keywords=['solve', 'optimization', 'problem', 'genetic', 'algorithm', 'GA', 'easy', 'fast', 'genetic-algorithm', 'combinatorial', 'mixed', 'evolutionary'],
     packages=setuptools.find_packages(),
     classifiers=[
@@ -27,7 +28,7 @@
         'numpy',
         'matplotlib',
         'joblib',
-        'OppOpPopInit'
+        'OppOpPopInit>=2.0.0'
         ]
 
     )