Matlab fmincon函数
Posted 流浪若相惜
tags:
篇首语:本文由小常识网(cha138.com)小编为大家整理,主要介绍了Matlab fmincon函数相关的知识,希望对你有一定的参考价值。
文章目录
函数功能
获取约束的非线性多变量函数的最小值
样式:
其中,c(x), ceq(x) 分别表示非线性的约束条件,而A, Aeq表示的是线性约束。
函数表达及用法
x = fmincon(fun,x0,A,b,Aeq,beq,lb,ub,nonlcon,options)
- fun
minimizes fun - 初始值
starts at x0,x0 can be a scalar, vector, or matrix. - 线性约束
attempts to find a minimizer x of the function described in fun subject to the linear inequalities Ax ≤ b,Aeqx = beq,and the range lb ≤ x ≤ ub - 非线性约束
the nonlinear inequalities c(x) or equalities ceq(x) defined in nonlcon
For example,x = fmincon(@myfun,x0,A,b,Aeq,beq,lb,ub,@mycon) where mycon is a MATLAB function such as function [c,ceq] = mycon(x) c = ... % Compute nonlinear inequalities at x. ceq = ... % Compute nonlinear equalities at x.
- 选项
Optimization options, specified as the output of optimoptions or a structure such as optimset returns
trust-region-reflective算法的说明 (梯度)
针对SpecialObjectiveGradient 将梯度计算加入目标函数中,以实现更快或更可靠的计算。将梯度计算作为条件化输出包含在目标函数文件中
function [f,g] = rosenbrockwithgrad(x)
% Calculate objective f
f = 100*(x(2) - x(1)^2)^2 + (1-x(1))^2;
if nargout > 1 % gradient required
g = [-400*(x(2)-x(1)^2)*x(1)-2*(1-x(1));
200*(x(2)-x(1)^2)];
end
注:
1.目标函数(标量)作为第一个输出,梯度(向量)作为第二个输出(应该是默认的)
2.使用 optimoptions 将 SpecifyObjectiveGradient 选项设置为 true。如果有的化,也将 SpecifyConstraintGradient 选项设置为 true。
fun = @rosenbrockwithgrad;
x0 = [-1,2];
A = [];
b = [];
Aeq = [];
beq = [];
lb = [-2,-2];
ub = [2,2];
nonlcon = [];
x = fmincon(fun,x0,A,b,Aeq,beq,lb,ub,nonlcon,options)
调用
options = optimoptions('fmincon','SpecifyObjectiveGradient',true);
输出结果
Local minimum found that satisfies the constraints.
Optimization completed because the objective function is non-decreasing in
feasible directions, to within the value of the optimality tolerance,
and constraints are satisfied to within the value of the constraint tolerance.
x =
1.0000 1.0000
Hessian矩阵应用说明
上例代码修改为
function [f, g, H] = rosenboth(x)
% Calculate objective f
f = 100*(x(2) - x(1)^2)^2 + (1-x(1))^2;
if nargout > 1 % gradient required
g = [-400*(x(2)-x(1)^2)*x(1)-2*(1-x(1));
200*(x(2)-x(1)^2)];
if nargout > 2 % Hessian required
H = [1200*x(1)^2-400*x(2)+2, -400*x(1);
-400*x(1), 200];
end
end
注:
1.使用 fmincon ‘trust-region-reflective’ 和 ‘interior-point’ 算法以及 fminunc ‘trust-region’ 算法来包含二阶导数。根据信息的类型和算法,可通过几种方法来包括 Hessian 矩阵信息。
2. 您还必须包含梯度(将 SpecifyObjectiveGradient 设置为 true,如果适用,还必须将 SpecifyConstraintGradient 设置为 true)以便包含 Hessian 矩阵
调用:
options = optimoptions('fminunc','Algorithm','trust-region',...
'SpecifyObjectiveGradient',true,'HessianFcn','objective');
适用于 fmincon 内点算法的 Hessian 矩阵. 该 Hessian 矩阵是拉格朗日函数的 Hessian 矩阵,其中 L(x,λ) 是
g 和 h 是向量函数,分别表示所有不等式和等式约束(指有界、线性和非线性约束),因此最小化问题的公式为
拉格朗日方程的 Hessian 矩阵公式为:
ssian 是 n×n 矩阵(稀疏或稠密),其中 n 是变量的数目。如果 hessian 很大并且非零项相对较少,请将 hessian 表示为稀疏矩阵,以节省运行时间和内存。lambda 是具有与非线性约束相关联的拉格朗日乘数向量的结构体:
lambda.ineqnonlin
lambda.eqnonlin
fmincon 计算结构体 lambda,并将其传递给您的 Hessian 函数。hessianfcn 必须计算上式中的总和。通过设置以下选项表明您提供了 Hessian 函数:
ptions = optimoptions('fmincon','Algorithm','interior-point',...
'SpecifyObjectiveGradient',true,'SpecifyConstraintGradient',true,...
'HessianFcn',@hessianfcn);
function Hout = hessianfcn(x,lambda)
% Hessian of objective
H = [1200*x(1)^2-400*x(2)+2, -400*x(1);
-400*x(1), 200];
% Hessian of nonlinear inequality constraint
Hg = 2*eye(2);
Hout = H + lambda.ineqnonlin*Hg;
将 hessianfcn 保存到 MATLAB 路径。为了完成此示例,包含梯度的约束函数为
function [c,ceq,gc,gceq] = unitdisk2(x)
c = x(1)^2 + x(2)^2 - 1;
ceq = [ ];
if nargout > 2
gc = [2*x(1);2*x(2)];
gceq = [];
end
调用:
fun = @rosenboth;
nonlcon = @unitdisk2;
x0 = [-1;2];
options = optimoptions('fmincon','Algorithm','interior-point',...
'SpecifyObjectiveGradient',true,'SpecifyConstraintGradient',true,...
'HessianFcn',@hessianfcn);
[x,fval,exitflag,output] = fmincon(fun,x0,[],[],[],[],[],[],@unitdisk2,options);
- Algorithm
返回值
[x,fval,exitflag,output,lambda,grad,hessian] = fmincon(___) additionally returns:
x— variable
fval— the value of the objective function fun
exitflag— describes the exit condition of fmincon,
output— a structure output with information about the optimization process.
lambda — Structure with fields containing the Lagrange multipliers at the solution x.
grad — Gradient of fun at the solution x.
hessian — Hessian of fun at the solution x. See fmincon Hessian.
有关拉格朗日算子到底是啥,可参考:如何理解拉格朗日乘子法?
梯度和海森矩阵请参考:牛顿法与拟牛顿法学习笔记
- exitflag
- output
- lambda
- Hessian
以上是关于Matlab fmincon函数的主要内容,如果未能解决你的问题,请参考以下文章