In the special case that the scaling factors r1, r2, ..., rN
all are equal, the formulas for α and f(α) can be simplified. |
  |
First note that in this case the expression for τ(q), |
p1qr1τ(q) +
p2qr2τ(q) + ... +
pNqrNτ(q) = 1 |
reduces to |
p1qrτ(q) +
p2qrτ(q) + ... +
pNqrτ(q) = 1 |
This can be solved explicitly for rτ(q) |
rτ(q) = 1/(p1q + p2q +
... + pNq) |
and so τ(q) is given by |
τ(q) = -Log(p1q + p2q +
... + pNq) / Log(r) |
  |
Next, the expression for α simplifies to |
α(q) = (rτ(q) (p1qLog(p1) + ...
+ pNqLog(pN))) / (rτ(q) Log(r) (p1q + ...
+ pNq)) |
hence to |
α(q) = (p1qLog(p1) + ...
+ pNqLog(pN)) / (Log(r) (p1q + ... + pNq)) |
  |
Finally, combining these f(α(q)) can be computed explicitly from the familiar formula |
f(α(q)) = q*α(q) + τ(q) |
The f(α) curve can be approximated by letting q range from some negative value, say q = -10, to some positive
value, say q = 10, in fairly small steps. This avoids the substantial headaches involved in trying to write f and an
explicit, even if only approximate, function of α. |
  |
For N = 4 and r = 0.5, we have |
f(α(q)) = q (p1qLog(p1) + ...
+ p4qLog(p4)) / (Log(0.5) (p1q + ... + p4q)) |
- Log(p1q + ... + p4q) / Log(0.5) |
|
|
|
|
|
|
|
|
|