Gauss全主元消去法.doc

Gaussian elimination with complete pivoting To keep the Gaussian elimination method smooth operation,We must ensure that the pivot elementandcan not too small in each operation.So?the selection of appropriate pivot ?elements?in every step of?elimination.,the?maximum absolute value?or?the major elements?as the pivot ?element，this improved Gaussian?elimination method?named?Gauss pivot?element elimination. There have 3 methods to select pivot element,Gaussian elimination with partial pivoting、scaled partial pivoting、complete pivoting.Through access to?some?data,I introduce the Gaussian elimination with complete pivoting briefly. Complete pivoting at the th step searches all the entries,to find the entry to the largest magnitude.Both row and column interchanges are performed to bring this entry to the pivot position. The specific steps： The th elimination,find all（i=k，k+1,,n; j=k,k+1n),the is the pivot element. Both row and column interchanges are performed to bring this entry to the pivot position.Move to (k,k).Until the n-1?step,?the original?equation into?the same?equations?on?triangular?solution: Backward substitution： Backward substitution is the solution?of triangular systems?.If，then Algorithm： function x=Gauss(A,b) B=[A,b]; n=length(A); x=zeros(n,2); for i=1:n x(i,1)=i; %x第一列数字i代表x（i） end for i=1:n-1 [m,p]=max(abs(B(i:n,i:n)));%求绝对值最大系数所在行号和列号 [l,q]=max(m); if m==0; disp(A奇异) return end if p(q)+i-1i %行变换 pm=p(q)+i-1; C(i:n+1)=B(i,i:n+1); B(i,i:n+1)=B(pm,i:n+1); B(pm,i:n+1)=C(i:n+1); end if q+i-1i %列变换 q=q+i-1; D(1:n)=B(1:n,q); B(1:n,q)=B(1:n,i); B(1:n,i)=D(1:n); E(1:2)=x(q,:);%调换解的顺序 x(q,:)=x(i,:); x(i,:)=E(1:2); end for k=i+1:n %消元 am=B(k,i)/B(i,i); B(k,i:n+1)=B(k,i:n+1)-am*B(i,i:n+1); end end x(n,2)=B(n,n+1)/B(n,n); for i=n-1:-1:1 %回代 xx(i+1:n)=x(i+1:n,2); x(i,2)=(B(i,n+1)-B(i,i+1:n)*xx(i+1:n))/B(i,i); end for i=1:n-1 % 调换解的顺序 [u,r]=min(x(i:n,1)); if r+i-1i r=r+i-1; xf(1:2)=x(i,1:2); x(i,1:2)=x(r,1:2); x(r,1:2)=xf(1:2); end end end For example P89，Exercis