lamda 是已知常数吧!
可以用newton raphson method
求解非线性联立方程

lamda 是拉格朗日乘子，也是未知的。是用拉格朗日乘子法求最大值的。这个能求么？

喔,你给的是微分后的方程.
Maximize the objective function : f(x,y)
subject to a constraint g(x,y)=c  or g(x,y)=<c
matlab有很多范例可参考.

谢谢，不过我写文章的时候，如果注明是用MATLAB的话，会有版权问题的。尤其是投外文的杂志。我看看那个开源的与MALAB类似的Scilab里面有没有类似的吧。谢谢

你可考虑free software Octave
http://www.gnu.org/software/octave/download.html

Numerical Methods Library就有OPT_LAGRANGE()
[XMIN, LAMBDAMIN, FMIN] = OPT_LAGRANGE(F, GRADF, G, JACG, X0)
computed the minimum of the function FUN subject to｀G(X) = 0＇
with the lagrange multiplier method.
Function GRADF defines the gradient vector of F.
Function G represents equality-constrained and
function JACG defines its jacobian matrix.
F and G accept a real vector input and return a real
vector. F, GRADF, G and JACG can also be inline object.

Parameters
F evaluated function.
GRADF F＇s gradient function.
G equality-constrained function : ＇G(X) = 0＇.
X0 initial point.

Returns
XMIN computed solution of min(FUN).
LAMBDAMIN vector of Lagrange multipliers on XMIN.
FX value of FUN(X) with X computed solution.