Find the Complete solution:Ax=b
x=xp+xn xp(particular solution) xn(solution in NULL space)
1.Ax=b->[A|b]->[R|d]
2.Let free variable =0,Rxp=d->xp
3.Rx=0;Let the free variable x0=1,the rest x=0->s1,s2...(special solution)
4.xn=c1*s1+c2*s2
5the answer x=xp+xn
Anylasis:
precondition:Ax=b is equal to Rx=d
The relationship bewteen r ,m and n
r:privot number m:row number n:column number
r<=min(m,n)
1.r=m=n
Shape: square
Solution number: one x=A^-1b xn=vector 0
Features:invertible matrix
2.r=m<n (full row rank)
Shape: short and fat
Solution number:infinite
use the method above to get the answers
Reasons:1.There are |m-n| excess variables in the equation.
Equation number<variable number,you can't confirm the answer,so the ansewer is infinite.
2.There are |m-n| excess vectors in the matirx.To get the given vectors,you have lots of ways
Features:1.free variable numbers are |n-m|
2.Nullspace!=0vector(have free variables)
3.the column spaceC(A) is R^m(Ax=b is always solvable)
3.r=n<m(full column rank)
1.Shape: tall and thin
2.Solution number: only or none.
Reasons:1.you can confirm the ansewer through n rows above
If the answers placed in the |m-n| equtions are not correct ,there are none .
Otherwise it is only one.
2.If d(Rx=d)don‘t have the same number rows
Features:1.There are |m-n| zero rows
2.Nullspace=0 vector(no free variable)
4.r<min(m,n)
Solution number: infinite or none
Features:matrix have zero rows and excess vector
5.picture different situations of R
r=m=n [I]
r=m<n [I F]
r=n<m: [I]
[0]
r<min(m,n): [I] [F]
[0][0]
question:为什么解是xp+xn
猜测一:A(xp+xn)=A(xp)+A(xn)=b+0=b
猜测二:向量空间具有像向量一样的性质,任意向量b可以看成先到达零向量,再由零向量到达向量b
猜测三:从答案来讲:多解的原因在于自由变量可以取不同的解,零向量的s1,s2相当于是自由列的基底,系数乘特解可以使得某自由变量变为任何数,可是仍漏了自由变量都为0
,其他数不为0的解,由此xp出现是必然。.从解的理解上xp+xn存在模拟全部解的可能