Python中的反向Cuthill-Mckee

Question

如何使用ScipyRCM返回的置换数组对原始稀疏矩阵进行重新排序并减少带宽

B = mmread('G22.mtx')
graph = csr_matrix(B)
aux2 = reverse_cuthill_mckee(graph,symmetric_mode=True)

其中“图”是无向图（对称矩阵）

Answer 1

我找到了答案，如果将来有人需要它：

B = mmread('G22.mtx')
graph = csr_matrix(B)
aux2 = reverse_cuthill_mckee(graph,symmetric_mode=True)


for i in range(len(aux2)):
    graph[:,i] = graph[aux2,i]
for i in range(len(aux2)):
    graph[i,:] = graph[i,aux2]