Question

我需要在networkx 2.1中生成一个随机的全连接有向图，以评估非对称旅行商问题的算法的性能。例如，生成具有100个节点的图，它们是完全连接的，边缘权重是随机分配的。该图是有向的（从节点i到节点j的边缘权重不一定等于从节点j到节点i的边缘权重）

想知道是否有一个networkx函数来生成这种有向图或一些示例代码以供参考。

我检查了networkx 2.1（https://networkx.github.io/documentation/networkx-2.1/reference/generators.html?highlight=generator#module-networkx.generators.directed）文档中的“定向生成器”部分，但是这些生成器似乎不符合我的要求。

Answer 1

Networkx并不完全具有这种生成器，因为它是非常特定的任务。因此，您应该手动构造该图。您可以创建完整的有向图：

export class RegistercustomerPage implements OnInit {
category= "product";
constructor(private geolocation: Geolocation){}

 ngOnInit() {
      this.loadMap();
  }
 loadMap(){

         this.geolocation.getCurrentPosition().then((resp) => {
            console.log(this.category);
            this.disaplyCategory();
        });
}
disaplyCategory(){
 alert(this.category)
}
}

，然后为每个图形边缘分配随机权重：

........................
            for book_data in data['items']:
                volume_info = book_data['volumeInfo']
                title = volume_info['title']
                genres = volume_info.get('categories')
                authors = volume_info.get('authors')
                description = volume_info.get('description')
                if not Books.objects.filter(title=title).exists():
                    book = Books.objects.create(title=title, description=description)

                    # Does authors exists in database?
                    existing_authors = Authors.objects.filter(author_name__in=authors)
                    existing_authors_names = {authors.author_name for authors in existing_authors}

                    # Create a list of missing authors
                    missing_authors = [
                        Authors(author_name=author_name)
                        for author_name in authors
                        if author_name not in existing_authors_names
                    ]
                    # Creating author before adding it to relation
                    if missing_authors:
                        missing_authors = Authors.objects.bulk_create(missing_authors)
                        print(missing_authors)
                        for m in missing_authors:
                            m.save()

                    # Adding to relation
                    book.authors.add(*existing_authors, *missing_authors)
..........................

因此您将准确获得所需的图形：

class Authors(models.Model):
    author_name = models.CharField(max_length=200)

    def __str__(self):
        return self.author_name


class Books(models.Model):
    title = models.CharField(max_length=300)
    description = models.TextField(blank=True, null=True)
    authors = models.ManyToManyField(Authors, blank=True)
    genres = models.ManyToManyField(Genres, blank=True)

    def __str__(self):
        return self.title

import networkx as nx
import random

N = 7
G = nx.complete_graph(N, nx.DiGraph())

Answer 2

有关如何执行此操作的几种选择。在这里，我仅使用itertools列出具有随机权重的所有加权边的列表。然后，我使用add_weighted_edges_from创建有向图。这样就不会太长，我只使用3个节点。

import networkx as nx
import itertools
import random

G = nx.DiGraph()
weighted_edge_list = [(u,v,random.random()) for u,v in itertools.permutations(range(3),2)]
G.add_weighted_edges_from(weighted_edge_list) 

G.edges(data=True)
> OutEdgeDataView([(0, 1, {'weight': 0.025851202944826346}), (0, 2, {'weight': 0.8067025754602839}), (1, 0, {'weight': 0.7729736390607577}), (1, 2, {'weight': 0.8724493159416196}), (2, 0, {'weight': 0.9049870220916731}), (2, 1, {'weight': 0.9636865700934618})])

如何在networkx 2.1中生成完全连接的有向随机图？

2 个答案: