Python求解条件最短路径问题：NetworkX

条件最短路径问题指带有约束条件的最短路径问题。例如顶点约束、必经点约束等。在本篇文章中，将通过案例，简要给大家说明一下，施加顶点、必经点约束的最短路径的求解。

 如上图所示，必经点为B、C、D，求解由A到E的一条最短路径。

1：创建无向图（上图）

import matplotlib.pyplot as plt  # 导入 Matplotlib 工具包
import networkx as nx  # 导入 NetworkX 工具包

gAnt = nx.Graph()  # 创建：空的 无向图
gAnt.add_weighted_edges_from([("A", 2, 3.3), ("A", 1, 8.1), ("A", 3, 8.8),
                              (1, "B", 3.9), (2, "B", 7.4), (2, "C", 7.7), (3, "D", 4.6), (3, "E", 5),
                              (4, "D", 0.72),
                              (4, "B", 1.0), (4, 5, 0.46), (5, "E", 1.44), (5, "D", 0.8),
                              ("C", 4, 0.94),
                              ("B", "C", 0.5), ("B", "D", 0.95),
                              ("D", "E", 1.2)
                              ])  # 向图中添加多条赋权边: (node1,node2,weight)

pos = {"A": (7, 15), "B": (6, 5), "C": (7, 4), "D": (5, 3), "E": (4, 1.5),
       1: (2, 8), 2: (9, 13), 3: (0, 6), 4: (6.5, 3), 5: (6.25, 2)}  # 指定节点在图中的位置
nx.draw(gAnt, pos, with_labels=True, alpha=0.8)
labels = nx.get_edge_attributes(gAnt, 'weight') # 边的权值，即两点间的距离
nx.draw_networkx_edge_labels(gAnt, pos, edge_labels=labels, font_color='r', font_size=8)  # 显示权值
nx.draw_networkx_nodes(gAnt, pos, nodelist=["A", "E"], node_color='yellow')  # 设置顶点颜色
nx.draw_networkx_nodes(gAnt, pos, nodelist=["B", "C", "D"], node_color='lime')  # 设置顶点颜色
plt.show()

 2：限制条件：多个必经点

解决方案：遍历从起点到终点的简单路径，求满足必经点条件的最短路径

minWPath4 = min([path  # 返回 key 为最小值的 path
    for path in nx.all_simple_paths(gAnt, "A", "E")  # gAnt 中所有起点为A、终点为E的简单路径
    if all(n in path for n in ("B", "C", "D"))],  # 满足路径中包括顶点 B,C,D
    key=lambda x: sum(gAnt.edges[edge]['weight'] for edge in nx.utils.pairwise(x)))  # key 为加权路径长度
lenPath = sum(gAnt.edges[edge]['weight'] for edge in nx.utils.pairwise(minWPath4))  # 求指定路径的加权路径长度
for i in range(0, len(minWPath4)-1):
    nx.draw_networkx_edges(gAnt, pos, edgelist=[(minWPath4[i], minWPath4[i+1])], edge_color='r', width=2.5)  # 设置边的颜色
print("\n问题4: 多个必经点的约束")
print("A 到 E 的最短加权路径: ", minWPath4)
print("A 到 E 的最短加权路径长度: ", lenPath)

输出结果：

多个必经点的约束
A 到 E 的最短加权路径:  ['A', 2, 'C', 'B', 'D', 'E']
A 到 E 的最短加权路径长度:  13.649999999999999

 Ref：Python数模笔记-NetworkX（3）条件最短路径