解线性方程组——高斯消去法

高斯消去法，通常包含两个过程，消去过程和回代过程，不同的高斯消去法其消去过程也不尽相同，一般的高斯消去法旨在将系数阵通过行变换，变成上三角矩阵，而列主元素的高斯消去法，在一般的基础上，每次行变换时将该列绝对值最大元素变换到主元位置上，而量化的列主元素高斯消去法则不进行行交换，通过一个l向量来存储列主元素所在行的位置，按照该顺序进行行变换，可以减少用于行交换的处理时间。

直接上代码，matlab编写：

1.高斯消去法

function x = gauss1(A,b) %A为系数阵，b为常数向量
    x1 = A;
    x2 = b;
    size_x1 = size(x1);
    n = size_x1(1); %n为方程数
    if A(1,1) == 0
        x = 'error';
        disp('error!!!');
        return;
    end %对于对角元为0不适用
    for i = 1:1:n
        col_i = x1(:,i); %取列向量
        if col_i(i) == 0
            x = 'error';
            disp('error!!!');
            return;
        end
        m_i = col_i / col_i(i); 
        for k1 = i+1:1:n
            x2(k1) = x2(k1) - m_i(k1) * x2(i);
            for k2 = 1:1:n
                x1(k1,k2) = x1(k1,k2) - m_i(k1) * x1(i,k2);
            end
        end
    end
    %回代过程
    res_x1 = x1;
    res_x2 = x2;
    x = zeros(size(res_x2));
    for i = n:-1:1
        sum = 0;
        if i == n
            x(n) = x2(n) / x1(n,n);
        else
            temp = res_x1(i,:) .* x';
            for j = 1:1:n
                sum = sum + temp(j);
            end
            x(i) = (x2(i) - sum) / x1(i,i);
        end
    end
    return;
end

2.列主元素高斯消去法

function x = gauss2(A,b)
    x1 = A;
    x2 = b;
    size_x1 = size(x1);
    n = size_x1(1);
    for i = 1:1:n
        col_i = x1(:,i);
        for ii = i+1:1:n
            if col_i(ii) > col_i(i)
                t1 = zeros(size(col_i));
                for jj = 1:1:n
                    t1(jj) = x1(i,jj);
                    x1(i,jj) = x1(ii,jj);
                    x1(ii,jj) = t1(jj);
                end
                t2 = x2(i);
                x2(i) = x2(ii);
                x2(ii) = t2;
            end
            col_i = x1(:,i);
        end
        if col_i(i) == 0
            x = 'error';
            disp('error!!!');
            return;
        end
        m_i = col_i / col_i(i);
        for k1 = i+1:1:n
            x2(k1) = x2(k1) - m_i(k1) * x2(i);
            for k2 = 1:1:n
                x1(k1,k2) = x1(k1,k2) - m_i(k1) * x1(i,k2);
            end
        end
    end
    res_x1 = x1;
    res_x2 = x2;
    x = zeros(size(res_x2));
    for i = n:-1:1
        sum = 0;
        if i == n
            x(n) = x2(n) / x1(n,n);
        else
            temp = res_x1(i,:) .* x';
            for j = 1:1:n
                sum = sum + temp(j);
            end
            x(i) = (x2(i) - sum) / x1(i,i);
        end
    end
    return;
end

3.量化的列主元素高斯消去法

function x = gauss3(A,b)
    size_A = size(A);
    n = size_A(1);
    l = 1:1:n;
    S = zeros(size(l));
    for i = 1:1:n
        for j = 1:1:n
            if abs(A(i,j)) > S(i)
                S(i) = abs(A(i,j));
            end
        end
    end
    for i = 1:1:n
        max_as = 0;
        col_i = A(:,i);
        for j = i:1:n
            if col_i(l(j)) / S(j) > max_as
                max_as = col_i(l(j)) / S(j);
                t = l(i);
                l(i) = l(j);
                l(j) = t;
            end
        end
        m_i = col_i / col_i(l(i));
        for k1 = i+1:1:n
            b(l(k1)) = b(l(k1)) - m_i(l(k1)) * b(l(i));
            for k2 = 1:1:n
                A(l(k1),k2) = A(l(k1),k2) - m_i(l(k1)) * A(l(i),k2);
            end
        end
    end
    x = zeros(size(b));
    for i = n:-1:1
        sum = 0;
        if l(i) == l(n)
            x(n) = b(l(n)) / A(l(n),n);
        else
            temp = A(l(i),:) .* x';
            for j = 1:1:n
                sum = sum + temp(j);
            end
            x(i) = (b(l(i)) - sum) / A(l(i),i);
        end
    end
    return;
end