Merge pull request #221 from ZXF1001/master

Parameters for the number of cpu cores used

docs/en/args.md

@@ -30,6 +30,7 @@ help(sko.AFSA.AFSA)
 | constraint\_eq  | tuple()    | equal constraint        |
 | constraint\_ueq | tuple()    | unequal constraint                  |
 | precision       | 1e\-7  | precision，int/float or list |
+| n_processes     | 0  | number of processes, 0 means use all cpu |
 
 
 ### GA_TSP
@@ -57,6 +58,7 @@ help(sko.AFSA.AFSA)
 | c1        | 0\.5 | cognitive parameter |
 | c2        | 0\.5 | social parameter |
 | constraint\_ueq | tuple()    | unequal constraint                  |
+| n_processes     | 0  | number of processes, 0 means use all cpu |
 
 ### DE
 
@@ -72,6 +74,7 @@ help(sko.AFSA.AFSA)
 | ub        | 1      | upper bound of variables |
 | constraint\_eq  | tuple()   | equal constraint                   |
 | constraint\_ueq | tuple()    | unequal constraint                 |
+| n_processes     | 0  | number of processes, 0 means use all cpu |
 
 ### SA
 

docs/zh/args.md

@@ -30,6 +30,7 @@ help(sko.AFSA.AFSA)
 | constraint\_eq  | 空元组    | 等式约束                   |
 | constraint\_ueq | 空元组    | 不等式约束                  |
 | precision       | 1e\-7  | 精准度，int/float或者它们组成的列表 |
+| n_processes     | 0  | 使用的核数, 0表示使用全部的核 |
 
 
 ### GA_TSP
@@ -57,6 +58,8 @@ help(sko.AFSA.AFSA)
 | c1        | 0\.5 | 个体记忆     |
 | c2        | 0\.5 | 集体记忆     |
 | constraint\_ueq | 空元组    | 不等式约束                  |
+| n_processes     | 0  | 使用的核数, 0表示使用全部的核 |
+
 
 ### DE
 
@@ -72,6 +75,8 @@ help(sko.AFSA.AFSA)
 | ub        | 1      | 每个自变量的最大值 |
 | constraint\_eq  | 空元组    | 等式约束                   |
 | constraint\_ueq | 空元组    | 不等式约束                  |
+| n_processes     | 0  | 使用的核数, 0表示使用全部的核 |
+
 
 ### SA
 

sko/DE.py

@@ -15,9 +15,10 @@ class DE(GeneticAlgorithmBase):
     def __init__(self, func, n_dim, F=0.5,
                  size_pop=50, max_iter=200, prob_mut=0.3,
                  lb=-1, ub=1,
-                 constraint_eq=tuple(), constraint_ueq=tuple()):
+                 constraint_eq=tuple(), constraint_ueq=tuple(),
+                 n_processes=0):
         super().__init__(func, n_dim, size_pop, max_iter, prob_mut,
-                         constraint_eq=constraint_eq, constraint_ueq=constraint_ueq)
+                         constraint_eq=constraint_eq, constraint_ueq=constraint_ueq, n_processes=n_processes)
 
         self.F = F
         self.V, self.U = None, None

sko/GA.py

@@ -14,8 +14,8 @@
 class GeneticAlgorithmBase(SkoBase, metaclass=ABCMeta):
     def __init__(self, func, n_dim,
                  size_pop=50, max_iter=200, prob_mut=0.001,
-                 constraint_eq=tuple(), constraint_ueq=tuple(), early_stop=None):
-        self.func = func_transformer(func)
+                 constraint_eq=tuple(), constraint_ueq=tuple(), early_stop=None, n_processes=0):
+        self.func = func_transformer(func, n_processes)
         assert size_pop % 2 == 0, 'size_pop must be even integer'
         self.size_pop = size_pop  # size of population
         self.max_iter = max_iter
@@ -133,6 +133,8 @@ class GA(GeneticAlgorithmBase):
         Max of iter
     prob_mut : float between 0 and 1
         Probability of mutation
+    n_processes : int
+        Number of processes, 0 means use all cpu
     Attributes
     ----------------------
     Lind : array_like
@@ -151,8 +153,8 @@ def __init__(self, func, n_dim,
                  prob_mut=0.001,
                  lb=-1, ub=1,
                  constraint_eq=tuple(), constraint_ueq=tuple(),
-                 precision=1e-7, early_stop=None):
-        super().__init__(func, n_dim, size_pop, max_iter, prob_mut, constraint_eq, constraint_ueq, early_stop)
+                 precision=1e-7, early_stop=None, n_processes=0):
+        super().__init__(func, n_dim, size_pop, max_iter, prob_mut, constraint_eq, constraint_ueq, early_stop, n_processes=n_processes)
 
         self.lb, self.ub = np.array(lb) * np.ones(self.n_dim), np.array(ub) * np.ones(self.n_dim)
         self.precision = np.array(precision) * np.ones(self.n_dim)  # works when precision is int, float, list or array
@@ -339,15 +341,18 @@ class RCGA(GeneticAlgorithmBase):
         The lower bound of every variables of func
     ub : array_like
         The upper bound of every variables of func
+    n_processes : int
+        Number of processes, 0 means use all cpu
     """
 
     def __init__(self, func, n_dim,
                  size_pop=50, max_iter=200,
                  prob_mut=0.001,
                  prob_cros=0.9,
                  lb=-1, ub=1,
+                 n_processes=0
                  ):
-        super().__init__(func, n_dim, size_pop, max_iter, prob_mut)
+        super().__init__(func, n_dim, size_pop, max_iter, prob_mut, n_processes=n_processes)
         self.lb, self.ub = np.array(lb) * np.ones(self.n_dim), np.array(ub) * np.ones(self.n_dim)
         self.prob_cros = prob_cros
         self.crtbp()

sko/PSO.py

@@ -61,6 +61,8 @@ class PSO(SkoBase):
         equal constraint. Note: not available yet.
     constraint_ueq : tuple
         unequal constraint
+    n_processes : int
+        Number of processes, 0 means use all cpu
     Attributes
     ----------------------
     pbest_x : array_like, shape is (pop,dim)
@@ -82,11 +84,11 @@ class PSO(SkoBase):
 
     def __init__(self, func, n_dim=None, pop=40, max_iter=150, lb=-1e5, ub=1e5, w=0.8, c1=0.5, c2=0.5,
                  constraint_eq=tuple(), constraint_ueq=tuple(), verbose=False
-                 , dim=None):
+                 , dim=None, n_processes=0):
 
         n_dim = n_dim or dim  # support the earlier version
 
-        self.func = func_transformer(func)
+        self.func = func_transformer(func, n_processes)
         self.w = w  # inertia
         self.cp, self.cg = c1, c2  # parameters to control personal best, global best respectively
         self.pop = pop  # number of particles

sko/tools.py

@@ -27,7 +27,7 @@ def set_run_mode(func, mode):
     return
 
 
-def func_transformer(func):
+def func_transformer(func, n_processes):
     '''
     transform this kind of function:
     ```
@@ -91,18 +91,24 @@ def func_warped(X):
 
         return func_warped
     elif mode == 'multithreading':
+        assert n_processes >= 0, 'n_processes should >= 0'
         from multiprocessing.dummy import Pool as ThreadPool
-
-        pool = ThreadPool()
+        if n_processes == 0:
+            pool = ThreadPool()
+        else:
+            pool = ThreadPool(n_processes)
 
         def func_transformed(X):
             return np.array(pool.map(func, X))
 
         return func_transformed
     elif mode == 'multiprocessing':
+        assert n_processes >= 0, 'n_processes should >= 0'
         from multiprocessing import Pool
-        pool = Pool()
-
+        if n_processes == 0:
+            pool = Pool()
+        else:
+            pool = Pool(n_processes)
         def func_transformed(X):
             return np.array(pool.map(func, X))