I am new to matlab coding and I would like to know how to plot a tour visiting all points in a minimum spanning tree (yes, TSP/TSM). I was given a set of points a matrix of 20x2 and I was able to find out the MST of these points and I need help figuring out how to plan a tour of these points of minimum possible distance?
my adj matrix for the MST is,
X_st =
0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0
0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0
1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 1
0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0
0 1 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0
0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0
0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 1 0
1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 1 0 0 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 1 0 0 0 0 0 1 0 1 0 0 0 1 0 0 0 0
0 0 0 0 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Obtained from kruskal algorithm to plot a MST of a complete graph.
My, neighbouring weighted matrix obtained from kruskal function is,
1 3
7 17
5 20
6 14
1 17
6 20
16 19
2 14
7 11
6 18
12 19
14 16
10 19
8 11
2 8
3 15
9 18
4 19
13 15
Any guidance would be much appreciated.
once you have extracted the points for the MST using krushkals algorithm u need to use f=figure then for each (x,y) point it has to be like f = f + plot(x1,y1,x2,y2,[options]) plot and the plot code should be surrounded by hold on hold off please let me know if the answer was helpful the complete snippet will be like
f = figure;
hold on
f = f + plot(x1,y1,x2,y2) //put this in a loop for all points
hold off
Related
I'm writing an optimal path planning algorithm for one of my projects, and am a bit stuck on how to generate the map. In order to test my algorithm, I want to generate random maps that the object will pass through with obstacles in the way. To generate the maps I am making a logical array with "1" for wall or obstacle and "0" when there is nothing in the way. I am having some issues placing the obstacles however.
For now, this is what I've written in MATLAB (though I think the core idea can be solved in any language):
% lenx = length in x-direction of map [m]
% leny = length in y-direction of map [m]
% numObs = number of obstacles
% obsLoc = 2D array with the central location(s) of the obstacle(s)
% [x1, y1; x2 y2; ...]
% obsSize = size of the obstacles (all same size) [m]
% assume the obstacles are square.
% res = resolution of the map [cell/m]
%
% Outputs:
% map = 2D logical array which translates into a map
%
%---------------------------------------------------------
function map = createMap(lenx, leny, numObs, obsLoc, obsSize, res)
%Making map of the given size
map = zeros(leny*res, lenx*res);
%Adding walls
map(:,1) = 1;
map(:, lenx*res) = 1;
map(1,:) = 1;
map(leny*res, :) = 1;
%Recalculating obstacle size for this grid
obsSize = obsSize - 1;
%Adding obstacles
for i = 1:1:numObs
map(obsLoc(i,1)*res-obsSize*res:obsLoc(i,1)*res+obsSize*res,...
obsLoc(i,2)*res-obsSize*res:obsLoc(i,2)*res+obsSize*res) = 1;
end
end
The obsSize parameter is representing the dimension of one of the sides of the square obstacle. Here is a sample solution I ran:
>> createMap(10,10,1,[5,5],1,1)
ans =
1 1 1 1 1 1 1 1 1 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 1 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 1
1 1 1 1 1 1 1 1 1 1
Looks good! But the issue begins for obstacle sizes and resolutions that are not 1 or 2...
>> createMap(10,10,1,[5,5],0.5,2)
ans =
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
>> createMap(10,10,1,[5,5],3,2)
ans =
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 1 1 1 1 1 1 1 1 1 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
For the first one, we would expect there to be a 1 somewhere close to the middle, but it's nowhere to be found. In the second one, we'd expect to find 6 ones in each direction (3m * 3cell/m = 6 cell), but we are getting 8 (It's also possible that this is not what my code is calculating, but this is what I'm hoping it does. Neither case is working, and I think it has to do with this line:
Recalculating obstacle size for this grid
obsSize = obsSize - 1;
I'm pretty sure I do have to recalculate the object size so it fits to the grid, but how can I do this for any grid and obstacle? Is there a way to scale that "-1" such that it acts according to the size of the grid/obstacle?
Let me know what you come up with, and if there's any other glaring issues in my code?
Thanks!
I want to create a MxN matrix as shown below:
[1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0;
0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0;
0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0 0 0 0;
0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0 0 0 0;
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1 0 0 0;
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 1 1]
I have window size, let's say, 5 and it moves 3 in every row. Is it possible to create such a matrix without using for loops? Or is there any optimum way to do it?
This is a one line solution:
reshape([reshape([ones(5,6);zeros(21,6)], 1,[]), ones(1,5)],[],7).'
note:
The desired matrix can be seen as concatenation of a [6, 5+21] matrix:
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
1 1 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
and a [1 ,5] matrix:
1 1 1 1 1
that reshaped to a [7 , 23] matrix.
Other solution using repelem + bsxfun + accumarray:
r = repelem (1:7,5);
c= bsxfun(#plus, ((1:5)-3).',3*(1:7));
out = accumarray([r(:) c(:)] ,1)
Indices of rows and columns of 1 s can be generated and accumarray can be used to create the desired matrix.
I am confusing about exponentiation of matrix in Galois Field 2. Assume that I have a matrix that is represented in Galois Field 2 (GF2). I want to take exponentiation of 30. That is,
A^30
In the matlab we have two way to do it.
First, We perform A power 30 and take mod of 2 (Note that A is a double matrix)
A1=mod(A^30,2)
Second way, we convert A to galois matrix and take exponentiation
A=gf(A,2)
A2=A^30
Actually, two way must same result. However, when I check A1 and A2. They have a different result. What is happen in here? Thanks
Let see my A matrix
A =
1 1 0 1 1 1 0 1 0 0 0 1 0 1 0
0 0 0 1 0 1 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 1 0 1 0 0 0 1
1 0 1 1 0 0 0 0 0 0 1 1 1 0 0
1 0 1 0 0 0 0 1 0 0 0 0 0 0 0
0 1 0 1 1 1 1 1 0 1 1 1 1 0 1
0 0 0 1 0 0 0 1 0 0 1 0 0 0 0
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
1 0 0 1 0 1 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1 0 0 1 1
1 0 1 0 0 0 0 0 0 0 1 0 0 0 1
0 1 0 0 0 0 0 1 0 0 0 0 0 0 0
1 0 1 1 1 1 1 1 1 0 1 1 1 0 1
0 1 0 0 0 1 0 0 0 1 0 0 0 1 0
1 0 0 1 0 1 0 0 0 1 0 0 0 1 1
Method1:
A1=
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0
Method 2
A2 =
1 1 0 1 1 1 0 1 0 0 0 1 0 1 0
0 0 0 1 0 1 0 0 0 0 0 0 0 1 0
0 0 0 0 0 0 0 0 1 0 1 0 0 0 1
1 0 1 1 0 0 0 0 0 0 1 1 1 0 0
1 0 1 0 0 0 0 1 0 0 0 0 0 0 0
0 1 0 1 1 1 1 1 0 1 1 1 1 0 1
0 0 0 1 0 0 0 1 0 0 1 0 0 0 0
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0
1 0 0 1 0 1 0 0 0 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 1 0 0 1 1
1 0 1 0 0 0 0 0 0 0 1 0 0 0 1
0 1 0 0 0 0 0 1 0 0 0 0 0 0 0
1 0 1 1 1 1 1 1 1 0 1 1 1 0 1
0 1 0 0 0 1 0 0 0 1 0 0 0 1 0
1 0 0 1 0 1 0 0 0 1 0 0 0 1 1
I need to construct the tech cycle constraint matrix Aa and the right side ba. The aim is building the technology cycle matrices in order to solve the scheduling linear problem constrained by Ax<=b. In this case -1 and +1 in A refers to the coefficients of the constraints of the problem such as starting times and precedences
TC = [1,2,3,4,6,7;1,2,5,4,6,7;2,5,6,7,0,0]; % Technology cycle
CT = [100,60,200,160,80,120;100,60,150,120,60,150;50,120,40,30,0,0]; % Cycle time
n_jb = size(TC,1); % number of jobs
n_op = sum(TC~=0,2); % number of operations for each job
N_op = sum(n_op); % total number of operations
c=1; % indice for constraints in Aa
Op=1; % counter for overall operation
n_tf = N_op - n_jb- sum(n_op==1); % number of job transfer between machines (also number of tech cycle constraint numbers)
Aa = zeros(n_tf,N_op); % Constraint matrx for tech cycle
ba = zeros(n_tf,1); % The right vector of the constraint function: Aa*x<=ba
for j=1:n_jb
if n_op(j)>1
for op=1:n_op(j)-1
Aa(c,Op)=-1;
Aa(c,Op+1)=1;
ba(c,1)=CT(j,op);
c=c+1;
Op=Op+1;
end
else
Op=Op+1;
end
Op=Op+1;
end
The output, like Aa is 3 """diagonal""" -1/+1 matrices:
-1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 -1 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 -1 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 -1 1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 -1 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 -1 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 -1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 -1 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 -1 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 -1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1
In order to be more precise in the following there is an image: showing the 3 different part of the matrix Aa. My question is: Is there a way to build the same this avoiding loops since A is not a 3x1 but will definitely become 30-50x1?
You can use diag to create the positive and negative ones. The second input to diag is to shift the diagonal to the side. In this case, 1 to the right.
Use cumsum to find the rows you want to remove. For n = [6, 6, 4], you want to remove the 6th, 12th and 16th row.
n = [6, 6, 4];
cols = sum(n);
A = -eye(cols) + diag(ones(cols-1,1), 1);
A(cumsum(n),:) = []
A =
-1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0
0 -1 1 0 0 0 0 0 0 0 0 0 0 0 0 0
0 0 -1 1 0 0 0 0 0 0 0 0 0 0 0 0
0 0 0 -1 1 0 0 0 0 0 0 0 0 0 0 0
0 0 0 0 -1 1 0 0 0 0 0 0 0 0 0 0
0 0 0 0 0 0 -1 1 0 0 0 0 0 0 0 0
0 0 0 0 0 0 0 -1 1 0 0 0 0 0 0 0
0 0 0 0 0 0 0 0 -1 1 0 0 0 0 0 0
0 0 0 0 0 0 0 0 0 -1 1 0 0 0 0 0
0 0 0 0 0 0 0 0 0 0 -1 1 0 0 0 0
0 0 0 0 0 0 0 0 0 0 0 0 -1 1 0 0
0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1 0
0 0 0 0 0 0 0 0 0 0 0 0 0 0 -1 1
For the following letter, I wish to add noise to it by changing 5 percent of the 1's into 0's. So far, I have the following code which turns them all into 0's. Can someone please point me in the right direction? Thank you!
letterA = [0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 1 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 0 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 0 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 0 ...
0 0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 0 ...
0 0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 0 ...
0 0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 0 ...
0 0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 0 ...
0 0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 0 ...
0 1 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 1 0 ...
0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0 0];
for i=1:numel(letterA)
if letterA(i)==1
letterA(i)=0;
end
end
disp(letterA)
try this:
letterA( letterA == 1 & rand(size(letterA)) <= 0.05 ) = 0;
In fact you could also do
letterA( rand(size(letterA)) <= 0.05 ) = 0;
which sets each element with probability of 5% to zero. The already zero elements are not affected. I think what causes confusion here is that you have to recognize that each element is independently handled from each other. It makes no difference if you do the first or the second version.
You can check it:
letterA = (rand(1e5,1) < 0.2); N1 = nnz(letterA);
letterA( rand(size(letterA)) <= 0.05 ) = 0;
(N1 - nnz(letterA))/N1
which gives values around 0.05, i.e. 5%. And it is not true what EitanT says, that it will flip at maximum 5%. It can be more than 5% or less, but on average it is 5%.
EitanTs version flippes exactly 5%, so which version to select depends on the application. For EitanT version the noise is correlated to the signal (because it is exact), which may or may not be what you want.
The basic approach is to find the indices of the 1's and count them, randomly pick a desired amount of indices out of them, and then operate on them:
one_flip_ratio = 0.05;
idx_ones = find(letterA == 1); %// Indices of 1's
flips = round(one_flip_ratio * numel(idx_ones)); %// Number of flips
idx_flips = idx_ones(randperm(numel(idx_ones), flips)); %// Indices of elements
letterA(idx_flips) = 0; %// Flip elements
This will flip 5% of the 1's to 0's.
Thanks for throwing out all these ideas, but eventually I came up with this. It will allow me to easily control both letter and background noise, which is what I intend to do. I'm just a novice, so this may not be the most efficient code, but it gets the job done! (I'm not looking for exactly 5%, the naked eye display value is what I'm more worried about.) PLEASE let me know how this can be improved! Thank you.
background_noise_intensity=0.05;
letter_noise_intensity=0.05;
for i=1:numel(letterA)
if letterA(i)==0
if rand < background_noise_intensity
letterA(i)=1;
end
elseif letterA(i)==1
if rand < letter_noise_intensity
letterA(i)=0;
end
end
end
noisy_letters=letterA;
reshaped_noisy_letters=reshape(noisy_letters,37,19)';
imshow(reshaped_noisy_letters);