efficient nested foor loop

a1=5:-1:-5;

b1=1:-1:-1;

for i=1:size(b1,2);

for j=1:size(a1,2);

Points(j,:,i) =[a1(1,j) b1(1,i)];% 2D points

end

end

F = reshape( permute(Points, [1 3 2]), [], 2);% 33 2D points

origin=[4.5 3.5];

Total_points=size(a1,2)*size(b1,2);

for i=1:Total_points;

distance(1,i)=sqrt(sum((F(i,:)-Total_points).^2));

original_current=100;

current_survived_first(1,i)=original_current/distance(1,i);

current_survived_second(1,i)=current_survived_first(1,i).*exp(-1.33)

Final_current(1,i)=original_current.*current_survived_second(1,i);

end

This set of matlab codes give me Final_current(1*33 double). Though I also want the same end result, I got it in expense of large intermediate matrices because I am considering all 33 points of interest in all steps. I want to write down nested for loop which first does calculation for 1st point and stores the end result in matrix 'Final_current', then it repeats all steps for remaining points and update the matrix 'Final_current'. So, after I define a1 and b1 I want to write down codes in a different way so that I will be able to store all my results in 'Final_current' matrix but none of intermediate matrices store results for all 33 points.

[Image Analyst]

There's no universe where 33 is considered "large." Or even a thousand times that.

Answer 1

[Andrei Bobrov]

a1 = 5:-1:-5;
b1 = 1:-1:-1;
[a b] = ndgrid(a1,b1);
F = [a(:) b(:)];
distance = sqrt(sum((F - size(F,1)).^2,2));
current_survived_first = 100./distance;
current_survived_second = current_survived_first*exp(-1.33);
Final_current = 100*current_survived_second;

OR

Final_current = 1e4./sqrt(sum(([repmat(a1.',numel(b1),1) reshape(repmat(b1,numel(a1),1),[],1)] - numel(a1)*numel(b1)).^2,2))*exp(-1.33);