In any channel, running at maximum information transfer, the residual entropy in the channel must be zero. Hence the power retained in the channel must always be normalized to one. Thus:
Power_in - Power_out = 1, but power is a square. So this is the hyperbolic flow constraint.
That does bring up a point. The adapted system is not a stable Shannon, it keeps entropy in the channel above zero, it transfers the map via connectivity, so we have:
Power_in-Power_out = Phi, I do believe. I think that adds asymmetry because:
phi^(1/2+n) + phi^(1/2 - n -1) is the hyperbolic form, but if the powers of Phi come from a sequence then the sequence is shifted by one to compose the hyperbolic:
Exp X 1/2 | 1/2+2 | 1/2+3
Exp 1/X 1/2-1 | 1/2 - 3 | 1/2 -4
So there is a charge shift. Now the result is still the hyperbolic flow model, but the connected network is offset. I think there is a Wythoff row that does this, let me check.
No comments:
Post a Comment