或
,给出
的共轭转置.
ConjugateTranspose
或
,给出
的共轭转置.
更多信息和选项
- ConjugateTranspose[m] 等价于 Conjugate[Transpose[m]]. »
可以通过输入
ct
或 \[ConjugateTranspose] 得到. - ConjugateTranspose[m] 也可以作为
给出,其中
通过输入
hc
或\[HermitianConjugate] 得到. - ConjugateTranspose[m] 的 StandardForm 和 TraditionalForm 格式为
- ConjugateTranspose[m,spec] 给出 Conjugate[Transpose[m,spec]]. »
范例
打开所有单元 关闭所有单元基本范例 (3)
范围 (11)
矩阵 (6)
m = (| | |
| ------ | ------ |
| a | b |
| 1 + 2I | 3 + 4I |)MatrixForm[ConjugateTranspose[m]]r = {{1.5, 2.2 I, 3.1 + 4.4I}};
c = Transpose[r]{r//MatrixForm, c//MatrixForm}Transpose[c]ConjugateTranspose[vec] 对项进行共轭,但不会改变 vec 的形状:
ConjugateTranspose[{1.5, 2.2 I, 3.1 + 4.4I}]ConjugateTranspose 适用于符号矩阵:
ct = ConjugateTranspose[{{a + b I, c + d I}}]使用 ComplexExpand 假设所有变量都是实数:
ComplexExpand[ct]ConjugateTranspose 保持单位矩阵不变:
ConjugateTranspose[IdentityMatrix[3]]//MatrixForms = SparseArray[{{i_, i_} -> 1, {i_, j_} /; Abs[i - j] == 1 -> i + j I}, {5, 5}]MatrixForm[s]ConjugateTranspose[s]MatrixForm[%]转置一个 SymmetrizedArray 对象:
sa = Symmetrize[RandomComplex[1 + I, {10, 10}], Antihermitian[{1, 2}]]ConjugateTranspose[sa]% === -sa数组 (5)
对秩为 3 的数组的前两层进行共轭转置,有效地将其视为向量矩阵:
a = Array[x, {2, 3, 2}];
a//MatrixFormConjugateTranspose[a]//MatrixForma = ArrayReshape[(1 + 2I)Range[24], {2, 3, 4}]ConjugateTranspose[a, {1, 3, 2}]ConjugateTranspose[a, {3, 2, 1}]ConjugateTranspose[a, {2, 1, 3}]使用 TwoWayRule 表示法进行转置:
a = RandomComplex[1 + I, {2, 3, 4, 5}];ConjugateTranspose[a, TwoWayRule[1, 4]] === ConjugateTranspose[a, {4, 2, 3, 1}]ConjugateTranspose[a, 2 4] === ConjugateTranspose[a, {1, 4, 3, 2}]a = ArrayReshape[Range[24]I, {1, 2, 3, 4}]ConjugateTranspose[a, 2 3]Dimensions[%]ConjugateTranspose[Array[a, {3, 3}], {1, 1}]应用 (10)
矩阵分解 (4)
m = RandomComplex[1 + I, {4, 4}];查找
的 QRDecomposition:
{q, r} = QRDecomposition[m];Chop[q.q^]q^.r - m//Chop计算矩阵
的 SchurDecomposition:
m = {{1.5, I}, {2.3I, -2 + .5I}};{q, t} = SchurDecomposition[m];Chop[q.ConjugateTranspose[q]]m - q.t.ConjugateTranspose[q]//Chop计算矩阵
的 SingularValueDecomposition:
m = {{1, I}, {2, 3 + 4I}};{u, Σ, v} = SingularValueDecomposition[m];MatrixForm /@ FullSimplify[{u.u, v.v}]m == u.Σ.ConjugateTranspose[v]//FullSimplifydims = {3, 5};
m = RandomComplex[1 + I, dims];{λ, vvecs} = Eigensystem[m.m];s = Sqrt[DeleteCases[Chop[λ], 0]]Σ = DiagonalMatrix[s, 0, dims];
MatrixForm[Chop[Σ]]v = Transpose[vvecs];u = m.Transpose[(Take[vvecs, Length[s]]/s)];u.u//Chopv.v//Chopu.Σ.v - m//Chop特数矩阵 (6)
埃尔米特矩阵服从
,反埃尔米矩阵服从
. 该矩阵为埃尔米特矩阵:
h = (| | | |
| --- | ------ | ------ |
| 1 | 2I | 3 |
| -2I | 4 | 5 + 6I |
| 3 | 5 - 6I | 7 |);h == ConjugateTranspose[h]使用 HermitianMatrixQ 进行验证:
HermitianMatrixQ[h]a = (| | | |
| -- | ------- | ------ |
| 0 | 1 | 2I |
| -1 | 0 | 3 + 4I |
| 2I | -3 + 4I | 0 |);a == -ConjugateTranspose[a]使用 AntihermitianMatrixQ 进行验证:
AntihermitianMatrixQ[a]u = {{(1/2), (ISqrt[3]/2)}, {(ISqrt[3]/2), (1/2)}};u.u == u.u == IdentityMatrix[2]使用 UnitaryMatrixQ 验证酉矩阵:
UnitaryMatrixQ[u]埃尔米特矩阵可酉对角化为
,其中
为是对角实数矩阵,
为酉矩阵. 验证以下矩阵是埃尔米特矩阵,然后对其进行对角化:
(h = {{1, 0, -2 * I}, {0, 1, 0}, {2 * I, 0, 3}})//MatrixFormConjugateTranspose[h] == hd = DiagonalMatrix[Eigenvalues[h]]v = FullSimplify[Normalize /@ Eigenvectors[h]]u = Transpose[v];
h == u.d.ConjugateTranspose[u]//FullSimplify反埃尔米特阵可酉对角化为
,其中
为对角矩阵且值为虚数,
为酉矩阵. 验证以下矩阵是反埃尔米矩阵,然后对其进行对角化:
(a = {{0, 1, -2 * I}, {-1, 0, 0}, {-2 * I, 0, 0}})//MatrixFormConjugateTranspose[a] == -ad = DiagonalMatrix[Eigenvalues[a]]v = FullSimplify[Normalize /@ Eigenvectors[a]]u = Transpose[v];
a == u.d.ConjugateTranspose[u]//FullSimplify酉矩阵本身可酉对角化为
,其中
为酉矩阵且
为对角矩阵,其项位于单位圆上. 验证以下矩阵是酉矩阵,然后对其进行对角化:
(u = {{(2/3), -(1/3), (2/3)}, {(2/3), (2/3), -(1/3)}, {-(1/3), (2/3), (2/3)}})//MatrixFormConjugateTranspose[u] == Inverse[u]{λ, vλ} = Eigensystem[u]Abs[λ]d = DiagonalMatrix[λ]v = Transpose[Normalize /@ vλ];u == v.d.ConjugateTranspose[v]//FullSimplify如果
,则矩阵
被称为正规矩阵. 正规矩阵是最通用的一种矩阵,它可以被酉对角化为
,其中
为对角矩阵,且
为酉矩阵. 所有埃尔米特矩阵
都是正规的,因为等式的两边都只是
:
TensorExpand[h.h == h.h == h.h, Assumptions -> h∈Matrices[{n, n}, Complexes, Hermitian[{1, 2}]]]TensorExpand[a.a == a.a == -a.a, Assumptions -> a∈Matrices[{n, n}, Complexes, Antihermitian[{1, 2}]]]酉矩阵是正规矩阵,因为在定义中代入
会在等式两边给出单位矩阵:
TensorReduce[u.u == u.u /. u -> Inverse[u], Assumptions -> u∈Matrices[{n, n}, Complexes]](n = {{5 + I, -2I}, {2, 4 + 2I}})//MatrixFormn.ConjugateTranspose[n] == ConjugateTranspose[n].n使用 NormalMatrixQ 进行验证:
NormalMatrixQ[n]像
这样的正规矩阵可使用 Eigensystem 进行酉对角化:
{λ, v} = Eigensystem[{n}](d = DiagonalMatrix[λ])//MatrixFormu = Transpose[Normalize /@ v];
UnitaryMatrixQ[u]n == u.d.ConjugateTranspose[u]属性和关系 (10)
ConjugateTranspose[m] 等价于 Conjugate[Transpose[m]]:
m = RandomComplex[1 + I, {2, 3}];ConjugateTranspose[m] - Conjugate[Transpose[m]]ConjugateTranspose 服从
:
a = RandomComplex[1, {3, 3}];
ConjugateTranspose[ConjugateTranspose[a]] == a对于兼容矩阵
和
,ConjugateTranspose 服从
:
a = RandomComplex[1, {3, 4}];
b = RandomComplex[1, {4, 5}];
(a.b) == b.a矩阵求逆与 ConjugateTranspose 互换,即
:
a = {{-(1/25), (12I/25), -(7I/25)}, {-(7I/25), (9/25), (1/25)}, {(13/75), -(31I/75), (16/75)}};Inverse[ConjugateTranspose[a]] == ConjugateTranspose[Inverse[a]]许多特殊矩阵根据其在 ConjugateTranspose 下的属性进行定义. 埃尔米特矩阵有
:
h = (| | |
| ----- | ----- |
| 1 | 2 + I |
| 2 - I | 3 |);
{ConjugateTranspose[h] == h, HermitianMatrixQ[h]}u = (| | |
| ----------- | ----------- |
| (Sqrt[3]/2) | (I/2) |
| (I/2) | (Sqrt[3]/2) |);
{ConjugateTranspose[u] == Inverse[u], UnitaryMatrixQ[u]}m = RandomComplex[1 + I, {3, 4}];h = m.m^;h == h^m = RandomComplex[1 + I, {4, 4}];h = m + m^;h == h^a = m - m^;
a == -a^Transpose[{{}}]结果不能再次为 {{}} ,因为维度 {1,0} 的排列是 {0,1} 且没有表达式有维度 {0,1}:
Dimensions[{{}}]ConjugateTranspose[a] 转置数组的前两层:
a = RandomComplex[10 + 5I, {5, 6, 7}];ConjugateTranspose[a] == ConjugateTranspose[a, {2, 1, 3}] == ConjugateTranspose[a, 1 <-> 2]ConjugateTranspose[a,perm] 返回维数为 Permute[Dimensions[a],perm] 的数组:
a = RandomInteger[10, {2, 3, 4, 5}];perm = RandomSample[Range[4]]Dimensions[ConjugateTranspose[a, perm]]Permute[Dimensions[a], perm]技术笔记
-
▪
- 矩阵的基本运算
相关指南
-
▪
- 矩阵运算 ▪
- 复数 ▪
- GPU 计算 ▪
- NVIDIA GPU 计算
历史
2004年引入 (5.1) | 在以下年份被更新:2025 (14.3)
文本
Wolfram Research (2004),ConjugateTranspose,Wolfram 语言函数,https://reference.wolfram.com/language/ref/ConjugateTranspose.html (更新于 2025 年).
CMS
Wolfram 语言. 2004. "ConjugateTranspose." Wolfram 语言与系统参考资料中心. Wolfram Research. 最新版本 2025. https://reference.wolfram.com/language/ref/ConjugateTranspose.html.
APA
Wolfram 语言. (2004). ConjugateTranspose. Wolfram 语言与系统参考资料中心. 追溯自 https://reference.wolfram.com/language/ref/ConjugateTranspose.html 年
BibTeX
@misc{reference.wolfram_2026_conjugatetranspose, author="Wolfram Research", title="{ConjugateTranspose}", year="2025", howpublished="\url{https://reference.wolfram.com/language/ref/ConjugateTranspose.html}", note=[Accessed: 13-September-2026]}
BibLaTeX
@online{reference.wolfram_2026_conjugatetranspose, organization={Wolfram Research}, title={ConjugateTranspose}, year={2025}, url={https://reference.wolfram.com/language/ref/ConjugateTranspose.html}, note=[Accessed: 13-September-2026]}