【Python3】非支配排序，帕累托前沿

最新推荐文章于 2024-08-23 11:00:00 发布

「已注销」

最新推荐文章于 2024-08-23 11:00:00 发布

阅读量2.5k

点赞数 2

分类专栏： NPprojrcts 文章标签： python

本文链接：https://blog.csdn.net/weixin_40775077/article/details/109897414

版权

本文介绍了如何使用Python3实现非支配排序的NSGA2遗传算法，探讨了该算法在多目标优化问题中的应用。

摘要由CSDN通过智能技术生成

test.py

import numpy as np
import matplotlib.pyplot as plt


class Utils:
    @staticmethod
    def is_dominate(obj_a, obj_b, num_obj, ):  # a dominates b
        if type(obj_a) is not np.ndarray:
            obj_a, obj_b = np.array(obj_a), np.array(obj_b)
        res = np.array([np.sign(k) for k in obj_a - obj_b])
        res_ngt0, res_eqf1 = np.argwhere(res <= 0), np.argwhere(res == -1)
        if res_ngt0.shape[0] == num_obj and res_eqf1.shape[0] > 0:
            return True
        return False


class Pareto:
    def __init__(self, pop_size, pop_obj, ):
        self.pop_size = pop_size
        self.pop_obj = pop_obj
        self.num_obj = pop_obj.shape[1]
        self.f = []
        self.sp = [[] for _ in range(pop_size)]
        self.np = np.zeros([pop_size, 1], dtype=int)
        self.rank = np.zeros([pop_size, 1], dtype=int)
        self.cd = np.zeros([pop_size, 1])

    def __index(self, i, ):
        return np.delete(range(self.pop_size), i)

    def __is_dominate(self, i, j, ):
        return Utils.is_dominate(self.pop_obj[i], self.pop_obj[j], self.num_obj)

    def __f1_dominate(self, )