Talk:Problem of Apollonius: Difference between revisions
→Meta error propagation
(→Meta error propagation: bout J) |
|||
Line 18:
:: Wait, not even for <code>[(0, -3), 2], [(0, 0), 1], [(0, 3), 2]</code>? --20:29, 15 September 2011 (UTC)
::: Yes. For that case, it starts at the 0,-3 point (which is the center of the first circle), finds that the excess volume is negative infinity, and so stops there. But the error for 0,-3 is 112 which is larger than the acceptable error volume. So it reports no result. I suppose the question is: what criteria should be used to determine when simplex has completed, if it's not the presence of excess volume? Alternatively, what value should be reported for this case? And, why? --[[User:Rdm|Rdm]] 20:54, 15 September 2011 (UTC)
| |||