|
|
import scipy as s
import scipy.sparse as ss
from scipy.sparse.construct import eye as speye
import scipy.sparse.linalg as ssl......A = ss.kron(Ay, speye(nx)) + ss.kron(speye(ny), Ax)这个“coo”稀疏形式矩阵，就是存一个坐标带一个值的那种如 ...... (2382, 2223) -64.0 (2382, 2381) -64.0 (2382, 2382) 256.0 (2382, 2383) -64.0 (2383, 2224) -64.0 (2383, 2382) -64.0 (2383, 2383) 256.0 (2383, 2384) -64.0 (2384, 2225) -64.0 (2384, 2383) -64.0 (2384, 2384) 256.0
如何对A求逆呢？
这是A的value [[ <2385x2385 sparse matrix of type '<type 'numpy.float64'>'with 11577 stored elements in Compressed Sparse Row format>]]
我尝试了一下 scipy.sprase.linalg.inv(A),发现根本行不通：AttributeError: 'matrix' object has no attribute 'format'
查了一下http://docs.scipy.org/doc/scipy/reference/generated/scipy.sparse.linalg.inv.html#scipy.sparse.linalg.inv发现貌似A可以使稀疏矩阵格式才对，我现在是搞不懂了，matlab能简单实现的稀疏矩阵求逆python会不行吗？求各位大佬帮我看看[source]
|
|
