| |
| | Admissible prime constellations |
 | | From these definitions it follows that s(x) is a piecewise-constant function (continuous on the right) that jumps from k-1 to k when x reaches l(k). |
 | | where C(0,a_2,...,a_k) is a non-zero constant for admissible constellations, to reduce the expected size of the first x that makes P(x;0,a_2,...,a_k) nonzero, it is crucial to either make C(0,a_2,...,a_k) as large as possible or to make k as small as possible. |
 | | It also appears that the absolute value of the error is bounded from above by the square root of k multiplied by a constant close to one (shaded area), at least for small values of k. |
| www.ieeta.pt /~tos/apc.html (1090 words) |
|