SoftMax for genetic Selection

sho/algo.py

@@ -68,15 +68,24 @@ def genetic(func, init, crossover, neighb, population_size, again):
     best_val = P_val[np.argmax(P_val)]
     i = 1
     while again(i, best_val, best_sol):
+        # Crossover and mutation
         new_sol = []
         for j in range(5*population_size):
             e1,e2 = np.random.randint(0,population_size,2)
             sol = neighb(crossover(P_sol[e1],P_sol[e2]))
             new_sol.append(sol)
+        # Selection
         new_val = [func(sol) for sol in new_sol]
-        P_i = np.random.choice(range(len(new_sol)), size = population_size, replace = False, p=[v/np.sum(new_val) for v in new_val])
+        # Probability folowing value order
+        order_val = np.argsort(new_val)
+        proba = np.zeros(len(new_val))
+        for j in range(len(new_val)):
+            proba[order_val[j]] = j
+        proba = np.exp(proba)/sum(np.exp(proba))
+        P_i = np.random.choice(range(len(new_sol)), size = population_size, replace = False, p=proba)
         P_sol = [new_sol[i] for i in P_i]
         P_val = [new_val[i] for i in P_i]
+        # Keeping best value
         if P_val[np.argmax(P_val)] > best_val:
             best_sol = P_sol[np.argmax(P_val)]
             best_val = P_val[np.argmax(P_val)]