Bubbels not connecting in plot command

Hello

I am using the following code and want the figure(2) to have bubbles connected with the 2 dashed lines(as you can see from the code). But for some reason I dont know, I am just getting the bubbles and no interconnection between the bubbels. This bothers me bcoz my trajectories are not visible clearly. Please help out. Please run the following code and see figure(2) plot if you cant understand what I mean.The code is:

clear all

clc

Ek=50;alpha=0.75;

for k=0:20,

figure(1)

stem(k,Ek);hold on

figure(2)

plot(k,3*(Ek)^0.5,'k--',k,-3*(Ek)^0.5,'k--o');hold on

Ek=(alpha^2)*Ek+1;

end

for j=1:100,

Xk=((50)^0.5)*randn(1,1);

for i=0:20,

figure(2)

plot(i,Xk,'b--o');hold on

Xk=alpha*Xk+randn(1,1);

end

end

Answer 1

[Adam]

Something more like this to replace your bottom section (i.e. starting from your for j = 1:100 loop) should give you joined up plots. The result looks incomprehensible to me, but I think it is achieving what you want.

i = 0:20;

for j = 1:100
   Xk = ((50)^0.5)*randn(21,1); Xk(2:end) = alpha*Xk(2:end)+randn(20,1);
   figure(2)
   plot(i,Xk,'b--o');hold on
end

There may be ways to further tidy up that code, but I was mostly focussed on just vectorising the Xk values to plot in one instruction rather than work out exactly what is happening to Xk!

[upinderjeet]

This has became to messy. Hehe. But thanks for ur help. So can you please tell me if I can get a plot with different colors. I mean ideally I would need one single color for one value of j and 99 other colors for remaining j's. I know we dont have so many colors but its ok if the colors repeat after 6 different colors. I wud atleast be able to distinguish the trajectories on the plot this way. And in this case my own previous code might still work bcoz we wont be worried about connecting the points. Please let me know

[Adam]

You can use:

plot(i,Xk,'--o','Color', [r g b]);hold on

to plot in the colour defined by [r g b]

So if you wanted 100 colours you could create them (though I agree they would not be distinguishable. Alternatively you can just define a cell array of 6 (or however many) RGB colours and in your plot command just plot as

plot(i,Xk,'--o','Color', colours{ mod( j, 6 ) + 1 ) };hold on

You can also use the other method of describing colours as e.g.

colours = {'r', 'g', 'b', 'k', 'y', 'm'};

which works with the above code too, but that is more limited as I think there are only ~8 colours defined that way and yellow is rather useless.

I think even that way though you have too much information on the plot to understand what is happening and from my attempts at joining the dots above it is quite possible that your trend could move about enough that one line of a given colour will cross over another, meaning that just 6 colours without trend lines will not give you that information.

[upinderjeet]

thanks adams