??? Subscripted assignment dimension mismatch

how do I fix this assignment error?

lambdaMJ = 7.2;                               
MJ_Flow = (30/60)*2;                                                                 
A = 100;                                  
B = 1;                                  
C = 70;                                      
D = 4;                                    
E = C + D;                       
F = A - C - D;     
MJ_start=0;                            
for a=1:100            
Num_cylces = 1;
XMJ=(poissrnd(lambdaMJ, A, Num_cylces));
lambdaMJ_sec(a) = mean(XMJ)/10;
% 
FMJ_cars_wait = @(T11,MJ_V11) (lambdaMJ_sec(a));
[T11,MJ_V11] = ode45(FMJ_cars_wait,[0,F],MJ_start);
R_G_MJ = MJ_V11(end:end);
%
FMJ_cars_wait_2 = @(T12,MJ_V12) (lambdaMJ_sec(a) - MJ_Flow);
[T12,MJ_V12] = ode45(FMJ_cars_wait_2,[F,A],R_G_MJ);
%
Major_end(a)=MJ_V12(end); 
MJ_start=MJ_V12(end);          
if MJ_start < 0            
    MJ_start=0;
else
    MJ_start=MJ_start; 
end
%   
Major1(:,a) = MJ_V11; 
end

??? Subscripted assignment dimension mismatch.

Error in ==> Untitled2 at 30
Major1(:,a) = MJ_V11;

Answer 1

[Cedric]

First:

 doc interp1

and once you understand roughly what it does and its input parameters, I guess that you could modify your code a little bit as follows (untested):

lambdaMJ = 7.2;                               
MJ_Flow = (30/60)*2;
A = 100;                                  
B = 1;                                  
C = 70;                                      
D = 4;                                    
E = C + D;                       
F = A - C - D;     
MJ_start=0;

% NEW
n            = 100 ;
step         = 0.5 ;
tGrid        = (0:step:F).' ;
lambdaMJ_sec = zeros(1, n) ;
Major_end    = zeros(1, n) ;
Major1       = zeros(length(tGrid), n) ;

for a = 1:n                               % Updated: upper bound.
Num_cylces = 1;
XMJ=(poissrnd(lambdaMJ, A, Num_cylces));
lambdaMJ_sec(a) = mean(XMJ)/10;
% 
FMJ_cars_wait = @(T11,MJ_V11) (lambdaMJ_sec(a));
[T11,MJ_V11] = ode45(FMJ_cars_wait,[0,F],MJ_start);
R_G_MJ = MJ_V11(end:end);
%
FMJ_cars_wait_2 = @(T12,MJ_V12) (lambdaMJ_sec(a) - MJ_Flow);
[T12,MJ_V12] = ode45(FMJ_cars_wait_2,[F,A],R_G_MJ);
%
Major_end(a)=MJ_V12(end); 
MJ_start=MJ_V12(end);          
if MJ_start < 0            
    MJ_start=0;
end                                              % Updated: no else clause.
%   
Major1(:,a) = interp1(T11, MJ_V11, tGrid);       % Updated: interpolate.
end

[harley]

thanks Cedric

[Cedric]

You're welcome. I didn't really get into the code .. is it correct that Num_cycles is constant and equal to 1 for all iterations of the loop? If so, it could be defined outside before the loop.

Answer 2

[Matt J]

Make MJ_V11 the same size as Major1(:,a)

[harley]

thanks Matt, should i preallocate? When i try this i still get the same error.

[Cedric]

I didn't fully look at your code, but often the issue in such context is that solutions outputted by different calls to ODE solvers are not expressed for the same set of values of the independent variable. We have therefore to interpolate all outputs on the same "grid".

[harley]

thanks Cedric, any suggestions?

[Image Analyst]

Matt, we went over all that in his original question http://www.mathworks.com/matlabcentral/answers/89151-how-to-find-a-mean-across-rows. I told him you can't stick a whole vector into a single element. Not sure why he abandoned that discussion, and asked all over again here.

[Cedric]

See my answer.