第四章 解线性方程组（Systems of Linear Equations）.doc

第四章 解線性方程組(Systems of Linear Equations) 在本章中包含 Matlab 的m-file nvgauss.m nvgauss_digit_k.m slgauss.m improve.m trigauss.m jacobi.m gseidel.m sor.m richardson.m steepest.m 7. conjgrad.m 將須要的m-file之檔案夾加入搜尋路徑中 path(d:\yeh\numerical, path) 註: 如果你有安裝Matlab Notebook 要執行下列input cells (綠色指令敘述) 之前必須先執行上面的cell – [path (…) ] 藍色的內容是Matlab [output cells] 1. nvgauss.m -- 自然的高斯消去法(Na?ve Gaussian Elimination) 顯示nvgauss.m的內容 type nvgauss.m function x= nvgauss( A, b, n) %to get an approximation solution of Ax=b, %by naive Gaussian Elimination, where A is n by n matrix %and obtain the solution by backward substitution % x=zeros(n,1); %initialize the solution vector A1 = A; b1=b ; for k = 1:n-1 if abs(A1(k,k)) eps for i= k+1:n m(i,k)=A1(i,k)/A1(k,k); for j = k+1:n A1(i,j)=A1(i,j)-m(i,k)*A1(k,j); end b1(i) = b1(i)-m(i,k)*b1(k); end %loop i else disp(Naive Gaussian Elimination fails) return end % if end %loop k % %to get the solution % if abs(A1(n,n)) eps x(n)=b1(n)/A1(n,n) ; %start the backward substitution for i = n-1:-1:1 s=b1(i) ; for j=i+1:n s= s-A1(i,j)*x(j); end %loop j x(i)=s/A1(i,i); end% loop i else disp(A(n,n) is zero, A is singular ) return end 例題 1: Solve the linear system. 6 x1 - 2 x2 + 2 x3 + 4 x4 = 16 12 x1 - 8 x2 + 6 x3 +10 x4 = 26 3 x1 -13 x2 + 9 x3 + 3 x4 = -19 -6 x1 + 4 x2 + x3 - 18 x4 = -34 準備必要的參數 A=[ 6 -2 2 4; 12 -8 6 10; 3 -13 9 3; -6 4 1 -18]; b = [16 26 -19 -34]; n = 4; x = nvgauss(A, b, n) x = 3 1 -2 1 這個