matlab联程multiple,Least squares minimization for multiple variables matlab

原文:

I need to find the value of tree variables: a, b and c, by finding a global minimum for least squares method. My function is as follows:

f = (1/a)*(asinh((Z(i)/b)^(1/c))^(-1)

where i is the index of vector Z. The vector Z has 9 values given in the task.

I also have a vector with 9 values for s.

The least squares method need to sum differences between values computed by function f with the values from vector s. It should look something like this:

((s(i)-f(i))/s(i))^2

I also have the boundaries for value a, b and c:

10e10>a>10e19, 10e-7>b>50, 10e-15>c>10.

I tried to use the lsqnonlin but I have no idea how to do it. I will appreciate any of your help!

I tried to do something like this:

function f=Fsigma(x, Z, sigma)

f=0;

for i=1:length(sigma)

f=f+((sigma(i)-((1/x(1,:))*(asinh((Z(i)/x(2,:))^(1/x(3,:)))^(-1))))/sigma(i))^2

end

end

and call this function in lsqnonlin like this:

Z= [1.49E+18 1.49E+19 1.49E+20 1.99E+15 1.99E+16 1.99E+17 1.49E+13 1.49E+14 1.49E+15];

sigma = [55.1705 79.1016 105.636 25.4809 40.8572 61.7238 12.8147 21.4054 34.8319];

a=linspace(10e10,10e19);

b=linspace(10e-7,50);

c=linspace(10e-15,10);

x=[a; b; c];

p=lsqnonlin(Fsigma(x,Z,sigma));

# Answer 1

4d350fd91e33782268f371d7edaa8a76.png

OK @Agata, I'll go through it in detail, so you can learn some basics.

First, how to hand functions to functions: function handles aka. the @-operator

fnc = @(x) Fsigma(x,Z,sigma);

fnc is an object -- in fact, it is a something that points at the function Fsigma. However, it even overshadows the additional inputs of Fsigma by declaring that its only input is x (this is called an anonymous function handle, indicated by the (), in between which you define the inputs that someone calling fnc may use. The other inputs to Fsigma in this line are the values of the variables that they have in this line.

boundaries

Boundaries should be provided as vectors:

% bounds

lb = [ 10e10;

10e-7;

10e-15];

ub = [ 10e19;

50;

10];

initial guess + optimization call if you read the docs of lsqnonlin, it requires an initial guess

% initial guess

x0 = ones(3,1);

% optimization call: x = lsqnonlin(fun,x0,lb,ub)

[p,fval] = lsqnonlin(fnc,x0,lb,ub)

Note to the cost function

lsqnonlin is better if you do not provide a summed cost but an array of errors (see the docs, again). So I adjust you Fsigma-function

function f = Fsigma(x, Z, sigma)

f = ((sigma-((1/x(1))./ asinh((Z./x(2)).^(1/x(3)))) )./sigma).^2;

end

If you want to have use a single output -- as most optimization algorithms require from their cost function -- you can use fmincon

Have a look at the full code

Z = [1.49E+18 1.49E+19 1.49E+20 1.99E+15 1.99E+16 1.99E+17 1.49E+13 1.49E+14 1.49E+15];

sigma = [55.1705 79.1016 105.636 25.4809 40.8572 61.7238 12.8147 21.4054 34.8319];

% bounds

lb = [ 10e10;

10e-7;

10e-15];

ub = [ 10e19;

50;

10];

% initial guess

x0 = ones(3,1);

% create an anyonymous function handle (using @(x)

fnc = @(x) Fsigma(x,Z,sigma);

% optimization call: x = lsqnonlin(fun,x0,lb,ub)

[p,fval] = lsqnonlin(fnc,x0,lb,ub);

% optimization call: x = fmincon(fun,x0,A,b,Aeq,beq,lb,ub)

fnc2 = @(x) sqrt(sum(fnc(x).^2));

[p2,fval2] = fmincon(fnc2,x0,[],[],[],[],lb,ub);

% cost function

function f = Fsigma(x, Z, sigma)

f = ((sigma-((1/x(1))./ asinh((Z./x(2)).^(1/x(3)))) )./sigma).^2;

end

PS: note that this community is not a coding service, so do your readings the next time!

  • 0
    点赞
  • 0
    收藏
    觉得还不错? 一键收藏
  • 0
    评论
该资源内项目源码是个人的课程设计、毕业设计,代码都测试ok,都是运行成功后才上传资源,答辩评审平均分达到96分,放心下载使用! ## 项目备注 1、该资源内项目代码都经过测试运行成功,功能ok的情况下才上传的,请放心下载使用! 2、本项目适合计算机相关专业(如计科、人工智能、通信工程、自动化、电子信息等)的在校学生、老师或者企业员工下载学习,也适合小白学习进阶,当然也可作为毕设项目、课程设计、作业、项目初期立项演示等。 3、如果基础还行,也可在此代码基础上进行修改,以实现其他功能,也可用于毕设、课设、作业等。 下载后请首先打开README.md文件(如有),仅供学习参考, 切勿用于商业用途。 该资源内项目源码是个人的课程设计,代码都测试ok,都是运行成功后才上传资源,答辩评审平均分达到96分,放心下载使用! ## 项目备注 1、该资源内项目代码都经过测试运行成功,功能ok的情况下才上传的,请放心下载使用! 2、本项目适合计算机相关专业(如计科、人工智能、通信工程、自动化、电子信息等)的在校学生、老师或者企业员工下载学习,也适合小白学习进阶,当然也可作为毕设项目、课程设计、作业、项目初期立项演示等。 3、如果基础还行,也可在此代码基础上进行修改,以实现其他功能,也可用于毕设、课设、作业等。 下载后请首先打开README.md文件(如有),仅供学习参考, 切勿用于商业用途。

“相关推荐”对你有帮助么?

  • 非常没帮助
  • 没帮助
  • 一般
  • 有帮助
  • 非常有帮助
提交
评论
添加红包

请填写红包祝福语或标题

红包个数最小为10个

红包金额最低5元

当前余额3.43前往充值 >
需支付:10.00
成就一亿技术人!
领取后你会自动成为博主和红包主的粉丝 规则
hope_wisdom
发出的红包
实付
使用余额支付
点击重新获取
扫码支付
钱包余额 0

抵扣说明:

1.余额是钱包充值的虚拟货币,按照1:1的比例进行支付金额的抵扣。
2.余额无法直接购买下载,可以购买VIP、付费专栏及课程。

余额充值