第二章 从几个例子来认识张量
2.1 前向变换和后向变换(Forward&Backward Transformations)
现在,我们有平面上的两组基:
![centering Old Basis &:{overrightarrow{e}_1,overrightarrow{e}_2}} \ \ \ New Basis &:{widetilde{overrightarrow{e}_1},widetilde{overrightarrow{e}_2}}}](data/attach/1904/so4vallh6ievk6h7e688jg416vcye8xh.jpg)
从Old Basis到New Basis的变换称为Forward,从New Basis到Old Basis的变换称为Backward。这两组基之间有如下关系:
![egin{matrix} widetilde{overrightarrow{e}_1}= 2overrightarrow{e}_1+1overrightarrow{e}_2 qquad qquad overrightarrow{e}_1=frac{1}{4}widetilde{overrightarrow{e}_1} +(-1)widetilde{overrightarrow{e}_2} \ widetilde{overrightarrow{e}_2} = -frac{1}{2}overrightarrow{e}_1+frac{1}{4}overrightarrow{e}_2 qquad qquad overrightarrow{e}_2=frac{1}{2}widetilde{overrightarrow{e}_1} +2widetilde{overrightarrow{e}_2} end{matrix}](data/attach/1904/8uzz6nza9k9v670ansh5w6ct81gjxr46.jpg)
所以,它们之间的变换矩阵如下:
![F=egin{bmatrix} 2 & -frac{1}{2} \ 1 & frac{1}{4} end{bmatrix} qquad qquad B=egin{bmatrix} frac{1}{4} & frac{1}{2} \ -1 & 2end{bmatrix}](data/attach/1904/57397oidhxtmn73tfrwav20wiz3lrkv8.jpg)
其中,F表示Forward transformation matrix,B表示Backward transformation matrix。将上式中的两个矩阵相乘有:
![FB=BF=I Rightarrow B=F^{-1}](data/attach/1904/15qbnpaljiggxnntfodmg3qkzxvtsuma.jpg)
将其中的数字抽象为数学符号,有:
![F=egin{bmatrix} F_{11} & F_{21} \ F_{12}& F_{22} end{bmatrix}](data/attach/1904/59o0yn59ykyhbm0g7u0mvdqkmusd8440.jpg)
扩展到n维向量空间,有:
![egin{matrix} widetilde{overrightarrow{e}_1} = F_{11}overrightarrow{e}_1+F_{21}overrightarrow{e}_2 + cdots + F_{n1}overrightarrow{e}_n \ widetilde{overrightarrow{e}_2} = F_{12}overrightarrow{e}_1+F_{22}overrightarrow{e}_2+ cdots + F_{n2}overrightarrow{e}_n \ vdots \ widetilde{overrightarrow{e}_n} = F_{1n}overrightarrow{e}_1+F_{2n}overrightarrow{e}_2+ cdots + F_{nn}overrightarrow{e}_n end{matrix} qquad qquad F = egin{bmatrix} F_{11} & F_{12} & cdots & F_{1n} \ F_{21} & F_{22} & cdots & F_{2n} \ vdots & vdots & &vdots\ F_{n1} & F_{n2} & cdots & F_{nn} \ end{bmatrix}](data/attach/1904/f2qxsk9p3jq1p8c1m2ldksk7hjzlhxa1.jpg)
所以,有:
![widetilde{overrightarrow{e}_i} = sum_{j=1}^{n}F_{ji}overrightarrow{e}_j](data/attach/1904/q7j4ikrji77v9bsm7mtnmyzb4bol76bt.jpg)
这里其实就是线性代数中的基变换和坐标变换,可以参考中国科学技术大学出版社的线性代数一书4.4节。不过这个变换是简单的,有线性代数基础的同学应该不难理解。
同理,对于Backward transformation matrix也有类似的结论:
![overrightarrow{e}_i= sum_{j=1}^{n}B_{ji}widetilde{overrightarrow{e}_j}](data/attach/1904/yd9k0wu4nvfehqiasjtfvygv6o1yhmtd.jpg)
由此可推得:
![egin{matrix} widetilde{overrightarrow{e}_j}= sum_{k=1}^{n}F_{kj}overrightarrow{e}_k \ overrightarrow{e}_i= sum_{j=1}^{n}B_{ji}widetilde{overrightarrow{e}_j} end{matrix} qquad Rightarrow qquad overrightarrow{e}_i= sum_{j=1}^{n}B_{ji}sum_{k=1}^{n}F_{kj}overrightarrow{e}_k=sum_{j=1}^{n}sum_{k=1}^{n}B_{ji}F_{kj}overrightarrow{e}_k](data/attach/1904/n9vjrq40ql9i20p3zhrqbji9nl7ughi4.jpg)
由左右两边相等可知:
![sum_{j=1}^{n}sum_{k=1}^{n}B_{ji}F_{kj}=left{egin{matrix} 1 &if i=k \ 0& if i
eq k end{matrix}
ight. =diag{1,1,cdots,1}=I](data/attach/1904/5rkz30c2akimmp2wj1o17otcyl83p69c.jpg)
将上述这种符号简化可得Kronecker Delta:
![delta_{ij}=left{egin{matrix} 1 &if i=k \ 0& if i
eq k end{matrix}
ight.](data/attach/1904/u1h5glxze6y6j02mfy9ah2dt9sn98rvf.jpg)
下一节,将会见到tensor的第一个例子。