hikaru1221
Aug22-11, 07:59 PM
Hi all,
I guess this is the appropriate place to put this. I'm currently co-working on a new channel model which has memory. My first task is to determine the channel capacity by simulation, as it seems very hard to find an explicit formula for this channel. Based on this channel model, I can simulate to get random outputs from any input. However, even with this, I do not know how to proceed next to compute the channel capacity.
C = lim _{n\rightarrow\infty} max \frac{1}{n}I(X^n;Y^n)
If I am to use directly this formula, i.e, gather a sufficient large number of statistics of x - input - and y - output, the number of computations to be made may be huge. Or perhaps that's my being inexperienced? Is there any way to get around this?
I guess this is the appropriate place to put this. I'm currently co-working on a new channel model which has memory. My first task is to determine the channel capacity by simulation, as it seems very hard to find an explicit formula for this channel. Based on this channel model, I can simulate to get random outputs from any input. However, even with this, I do not know how to proceed next to compute the channel capacity.
C = lim _{n\rightarrow\infty} max \frac{1}{n}I(X^n;Y^n)
If I am to use directly this formula, i.e, gather a sufficient large number of statistics of x - input - and y - output, the number of computations to be made may be huge. Or perhaps that's my being inexperienced? Is there any way to get around this?