Looping and indexing with global variables

Question

Adam Stephens 2020-6-12

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/547371-looping-and-indexing-with-global-variables

回答： Divya Gaddipati 2020-6-15

I am running a code I found attached to a research paper for a project I am working on and am running into an issue where the last loop of the main call function breaks the code. The for loop iterates over global variable s and runs other functions based on s. The function genalgodiffr runs, but breaks when calling mainfun2 and says that the index in position 1 exceeds array bounds when trying to index arrays to determine the value of diff_c. mainfun2 runs, and it will print the value of s = 1, so I know the code is running up until this line. I have also tried making s a local variable and making it a function paramater, but to no avail. Can anyone offer some tips on what to look at from here? Let me know if you need to see anything more of the code. Thank you.

%These are the loops where all variables used in diff_c are defined
    for i=0:30
        rb(i+1,1)= rb_max-i*.001638;
        ra(i+1,1)= rb(i+1,1)*tv(i+1,1);
        xa(i+1,1)= 2*(ra(i+1,1)-ra_min)*tand(alp)+.003;
        xb(i+1,1)= 2*(rb_max-rb(i+1,1))*tand(alp);
        bv(i+1,1)= (atand((ra(i+1,1)-rb(i+1,1))/l_h));
        th_a(i+1,1)= 180+2*(atand((ra(i+1,1)-rb(i+1,1))/l_h)) ;
        th_b(i+1,1)= 180-2*(atand((ra(i+1,1)-rb(i+1,1))/l_h));
        th_sb(i+1,1)= th_b(i+1,1)*0.8;
    end
    
    for i=0:30
        k(i+1,1)= k0*((rb(i+1,1)/rb(end-1,1))^2);
        k1(i+1,1)= k(i+1,1)* ((tand(alp+ph))/tand(alp)); 
        k2(i+1,1)= k(i+1,1)*((tand(alp-ph))/tand(alp));
        p(i+1,1)= sqrt(((1.5-(f*tand(alp)))^2)-(1.5-(f*tand(alp)))+(2*k(i+1,1)*(1+(f*cotd(alp)))));
        c_b(i+1,1)=tv_l(i+1,1)*c_a;
        om(i+1,1)= sqrt((k1(i+1,1)-p(i+1,1))*(p(i+1,1)-k2(i+1,1))*cosd(alp+ph)*cosd(alp-ph))/(p(i+1,1)*cosd(ph));
        zb(i+1,1)= 0.0174*th_b(i+1,1)+((k1(i+1,1)-k2(i+1,1))/(2*(p(i+1,1)-k2(i+1,1))))*(((sinh(om(i+1,1)*th_sb(i+1,1)*0.0174))/(om(i+1,1)))-th_sb(i+1,1));
        za(i+1,1)= 0.0174*th_a(i+1,1);
    end
  
    for i=0:29
        diff_ra(i+1,1)=ra(i+2,1)-ra(i+1,1);
        diff_xa(i+1,1)=2*diff_ra(i+1,1)*tand(alp); 
        diff_rb(i+1,1)=rb(i+2,1)-rb(i+1,1);
        diff_xb(i+1,1)=2*diff_rb(i+1,1)*tand(alp); 
        diff_za(i+1,1)=za(i+2,1)-za(i+1,1);
        diff_zb(i+1,1)=zb(i+2,1)-zb(i+1,1);
        diff_k(i+1,1)=k0*2*(rb(i+1,1)/(rb(end-1,1)^2))*diff_rb(i+1,1); 
        diff_th_a(i+1,1)=2*(cosd(bv(i+1,1))^2)* ((diff_ra(i+1,1)*((1/l_h)-(1/(tv(i+1,1)*l_h))))+(ra(i+1,1)/(tv(i+1,1)*tv(i+1,1)*l_h)));
        diff_th_b(i+1,1)= -diff_th_a(i+1,1) ;
        diff_cb(i+1,1)= -c_a/(tv(i+1,1)^2);
        diff_p(i+1,1)= (1+(f*cotd(alp))*diff_k(i+1,1))/p(i+1,1);
    end
    
%This is the for loop with global variable s
for s = 1:30
       [n,fval2]= genalgodiffr;
       fvaldiff(s+1,1)=fval2;
       v1 = ka(s,1)+ n(1)*(tv(s+1,1)-tv(s,1));
       v2 = r(s,1)+n(2)*(tv(s+1,1)-tv(s,1)); 
       v3 = deL(s,1)+n(3)*(tv(s+1,1)-tv(s,1)); 
       ka(s+1,1) = ka(s,1) + n(1)*(tv(s+1,1)-tv(s,1)); 
       r(s+1,1)= r(s,1)+n(2)*(tv(s+1,1)-tv(s,1)); 
       deL(s+1,1) = deL(s,1)+n(3)*(tv(s+1,1)-tv(s,1));
end
%This is the function genalgodiffr() that will run. mainfun2 is called with ga()
function [y,fval]= genalgodiffr()
    global tv s;
    nvars=  3;
    rng('default');
    lb= [70  .0010 1]./(tv(s+1,1)-tv(s,1));
    ub= [150  .0015 2]./(tv(s+1,1)-tv(s,1));  
    options= gaoptimset('Generation',150,'FitnessLimit',0,'PopulationSize',70,'CrossoverFraction',.6);
    [y,fval]= ga(@mainfun2, nvars,[],[],[],[],lb,ub,[],options);
end
%This is mainfun2
function diff_w = mainfun2(y) 
    global  fx n cs deL xa xb za zb kb alp d ra c_a diff_xb diff_za diff_zb diff_ra diff_xa diff_ca tv v1 v2 v3 s;
    deL(2,1)
    deL(s+1,1)
    deL(s,1)
    xb(s+1,1)
    xb(s,1)
    za(s+1,1)
    zb(s+1,1)
    m(1,1)=cs(4);
    a=(kb*deL(s+1,1)*(((180*(xb(s+1,1)-xb(s,1)))/(deL(s+1,1)-deL(s,1)*3.14*.0385)))*za(s+1,1))/(d*zb(s+1,1));
    c=((c_a*za(s+1,1)*(cotd(alp)))/(2*ra(s+1,1)));
    
    if s==1
        ka(1,1) =cs(1);
        r(1,1)  =cs(3);
        deL(1,1)=cs(5);
    else
        ka(s,1) = v1+n(1)*(tv(s+1,1)-tv(s,1));
        r(s,1)  = v2+n(2)*(tv(s+1,1)-tv(s,1));
        deL(s,1)= v3+n(3)*(tv(s+1,1)-tv(s,1));
    end
   
    if s==1
        di=(ka(s,1)*xa(s,1));
    else
      for l=2:s
        fx(l,1)=ka(l,1)*(xa(l,1)-xa(l-1,1));
      end
      di=sum(fx);
    end
    e=4*m(1,1)*r(s,1)*(y(2)*(tv(s+1)-tv(s)))/(xa(s+1,1)-xa(s,1));
    w= sqrt((a+c+di)/e)+3400;
    diff_a= y(3)*(xb(s+1,1)-xb(s,1))/(y(3)*(tv(s+1,1)-tv(s,1)))*kb*za(s+1,1)/(d*zb(s+1,1))+(((deL(s,1)*3.14/180)*y(3)*(sec(atan((xb(s+1,1)-xb(s,1))*180/(y(3)*3.14*(tv(s+1,1)-tv(s,1))))))^2)*(kb*za(s+1,1)/(zb(s+1,1)*d)))+diff_za(s,1)*kb*(((deL(s,1)+y(3)*(tv(s+1,1))-tv(s,1)))*((xb(s+1,1)-xb(s,1))/(y(3)*(tv(s+1,1)-tv(s,1)))))/(zb(s+1,1)*d)-(diff_zb(s,1)*kb*deL(s,1))-(diff_zb(s,1)*kb*(((deL(s,1)+y(3)*(tv(s+1,1)-tv(s,1)))*((xb(s+1,1)-xb(s,1))/((y(3)*(tv(s+1,1)-tv(s,1)*za(s+1,1))/(d*(zb(s+1,1))^2)-(kb*((deL(s,1)+y(3)*(tv(s+1,1)-tv(s,1)))*diff_xb(s,1)*((deL(s,1)+y(3)*(tv(s+1,1)-tv(s,1)))*3.14*.0385))))*za(s+1,1))/((d*zb(s+1,1)*(180*(xb(s+1,1))^2)))))));
    
    %This is the line where the code breaks
    diff_c = ((diff_ca(s,1)*za(s+1,1)*cotd(alp))/(2*ra(s+1,1)))+ ((diff_za(s,1)*c_a*cotd(alp))/(2*ra(s+1,1)))-((diff_ra(s,1)*c_a*za(s+1,1)*cotd(alp))/(2*(ra(s+1,1)))^2);
    diff_di=(y(1)*xa(s,1))+(diff_xa(s,1)*ka(s,1));
    diff_e=(y(2)*4*m(1,1)*(y(2)-r(s,1))/(xa(s+1)-xa(s)))+(y(2)*4*m(1,1)*r(s,1)*(secd(atand(y(2)-r(s,1))/(xa(s+1)-xa(s))))^2);
    diff_w= ((((a+c+di)*diff_e)/(e^2))-((diff_a+diff_c+diff_di)/e))/((2)*w);
end

2 个评论
显示无隐藏无

dpb 2020-6-13

That's a mess for sure...w/o being able to try to run the whole thing, not much chance anybody can just look at part of the code and have a clue -- particularly with all the globals all over everywhere -- they confuse things immensely as far as knowing "who's who in the zoo".

At least run the code and paste the complete error message in context to give a few hints...

Adam Stephens 2020-6-14

在 MATLAB Online 中打开

CVT_Genetic_Algorithm_6-4.zip

Gotcha, here is the full error message. I have also attached a zip file with all the code so far.

Index in position 1 exceeds array bounds.
Error in mainfun2 (line 37)
    diff_c = ((diff_ca(s,1)*za(s+1,1)*cotd(alp))/(2*ra(s+1,1)))+
    ((diff_za(s,1)*c_a*cotd(alp))/(2*ra(s+1,1)))-((diff_ra(s,1)*c_a*za(s+1,1)*cotd(alp))/(2*(ra(s+1,1)))^2);
Error in createAnonymousFcn>@(x)fcn(x,FcnArgs{:}) (line 11)
fcn_handle = @(x) fcn(x,FcnArgs{:});
Error in makeState (line 47)
            firstMemberScore = FitnessFcn(state.Population(initScoreProvided+1,:));
Error in galincon (line 17)
state = makeState(GenomeLength,FitnessFcn,Iterate,output.problemtype,options);
Error in ga (line 401)
            [x,fval,exitFlag,output,population,scores] = galincon(FitnessFcn,nvars, ...
Error in genalgodiffr (line 9)
    [y,fval]= ga(@mainfun2, nvars,[],[],[],[],lb,ub,[],options);
Error in maincall
Caused by:
    Failure in initial user-supplied fitness function evaluation. GA cannot continue.
 

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Divya Gaddipati 2020-6-15

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/547371-looping-and-indexing-with-global-variables#answer_451517

在 MATLAB Online 中打开

Hi,

Seems like the variable "diff_ca" is empty. Try initializing it to an array of zeros first time the function mainfun2 is called.

if isempty(diff_ca)
    diff_ca = zeros(n,1);
end

Also, make sure to check the sizes of all the variables on which you are performing operations.

Hope this helps!

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

Looping and indexing with global variables

2 个评论
显示无隐藏无

回答（1 个）

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

Looping and indexing with global variables

2 个评论 显示 无隐藏 无

回答（1 个）

0 个评论 显示 -2更早的评论隐藏 -2更早的评论

另请参阅

类别

标签

产品

版本

Community Treasure Hunt

2 个评论
显示无隐藏无

0 个评论
显示 -2更早的评论隐藏 -2更早的评论