python--代码实现差分进化算法（DE）并对测试函数进行测试及其边界处理.rar

# -*- coding: cp936 -*- import numpy as np import matplotlib.pyplot as plt import math import random # 采用DE/best/1/bin的差分进化策略 # 测试函数 def object_function(x): f = 0 for c in range(0, len(x)): f = f+(-x[c]*math.sin(math.sqrt(abs(x[c])))) return f # 种群初始化 def initialization(zq): np_list1 = [] for a in range(0, zq): x_list = [] for j in range(0, len_x): x_list.append(value_down_range + random.random() * (value_up_range - value_down_range)) np_list1.append(x_list) return np_list1 # 列表相减 def subtraction(a_list, b_list): a = len(a_list) new_list = [] for c in range(0, a): new_list.append(a_list[c] - b_list[c]) return new_list # 列表相加 def addition(a_list, b_list): a = len(a_list) new_list = [] for c in range(0, a): new_list.append(a_list[c] + b_list[c]) return new_list # 列表的数乘 def multiplier(a, b_list): b = len(b_list) new_list = [] for c in range(0, b): new_list.append(a * b_list[c]) return new_list # 变异 def mutation(np_list1): v_list1 = [] for gen in range(1, generation+1): d = math.exp(1-generation/(generation+1-gen)) f = F0*math.pow(2, d) for c in range(0, NP): r1 = random.randint(0, NP - 1) while r1 == c: r1 = random.randint(0, NP - 1) r2 = random.randint(0, NP - 1) while r2 == r1 | r2 == c: r2 = random.randint(0, NP - 1) r3 = random.randint(0, NP - 1) while r3 == r2 | r3 == r1 | r3 == c: r3 = random.randint(0, NP - 1) v_list1.append(addition(np_list1[r1], multiplier(f, subtraction(np_list1[r2], np_list1[r3])))) return v_list1 # 交叉 def crossover(np_list1, v_list1): u_list1 = [] for c in range(0, NP): vv_list = [] for j in range(0, len_x): if (random.random() <= CR) | (j == random.randint(0, len_x - 1)): if (v_list[c][j] < value_down_range) | (v_list[c][j] > value_up_range): v_list[c][j] = random.random() * (value_up_range - value_down_range) + value_down_range vv_list.append(v_list1[c][j]) else: if (np_list[c][j] < value_down_range) | (np_list[c][j] > value_up_range): np_list[c][j] = random.random() * (value_up_range - value_down_range) + value_down_range vv_list.append(np_list1[c][j]) u_list1.append(vv_list) return u_list1 # 选择 def selection(u_list1, np_list1): for c in range(0, NP): if object_function(u_list1[c]) <= object_function(np_list1[c]): np_list1[c] = u_list1[c] else: np_list1[c] = np_list1[c] return np_list1 # 主函数 # 参数 NP = 150 F0 = 0.4 CR = 0.5*(1+random.uniform(0, 1)) generation = 5000 len_x = 30 value_up_range = 500 value_down_range = -500 np_list = initialization(NP) min_x = [] min_f = [] xx = [] xx2 = [] for i in range(0, NP): xx.append(object_function(np_list[i])) min_f.append(min(xx)) min_x.append(np_list[xx.index(min(xx))]) for m in range(1, generation+1): v_list = mutation(np_list) u_list = crossover(np_list, v_list) np_list = selection(u_list, np_list) xx1 = [] for n in range(0, NP): xx1.append(object_function(np_list[n])) min_f.append(min(xx1)) min_x.append(np_list[xx1.index(min(xx1))]) min_ff1 = min(min_f) xx2.append(min_ff1) print('迭代次数：%d 当前最小值：%f' % (m, min_ff1)) # 输出 min_ff = min(min_f) min_xx = min_x[min_f.index(min_ff)] print('最小值点：%s' % min_xx) print('测试函数的最小值：%f' % min_ff) # 画图 x_label = np.arange(0, generation, 1) y_label = xx2 plt.rcParams['axes.unicode_minus'] = False plt.plot(x_label, y_label, color='blue') plt.xlabel('迭代次数') plt.ylabel('log10(f-fmin)') plt.savefig(r'S:\pycharm ONE\test4\20.png') plt.show()