tag:blogger.com,1999:blog-3891434218564545511.post101224518324954394..comments2024-03-28T19:56:42.305-05:00Comments on Alexander Pruss's Blog: The Tammes problemAlexander R Prusshttp://www.blogger.com/profile/05989277655934827117noreply@blogger.comBlogger2125tag:blogger.com,1999:blog-3891434218564545511.post-9016691673661186012017-01-23T11:29:16.444-06:002017-01-23T11:29:16.444-06:00I improved the greedy optimization step by taking ...I improved the greedy optimization step by taking the six nearest neighbors of each point, and then trying to move the point to the circumcenter of each of the 20 triangles formed by these neighbors (I think--haven't written down a formal proof, though--that at the optimum, each point will have to be at such a circumcenter). And then repeating this process 50 times. <br /><br />One can also specify -g on the tammes commandline in order to start with a golden angle spiral instead of a random arrangement.Alexander R Prusshttps://www.blogger.com/profile/05989277655934827117noreply@blogger.comtag:blogger.com,1999:blog-3891434218564545511.post-4440616337934857972017-01-23T08:46:10.876-06:002017-01-23T08:46:10.876-06:00I managed to get about a 2% improvement by adding ...I managed to get about a 2% improvement by adding 50 iterations of a cleanup stage. At each cleanup stage, I go through the particles, and I add a bit of greedy optimization by first shifting each particle from its nearest neighbor (unless there is a tie) and then from the midpoint of its two nearest neighbors.Alexander R Prusshttps://www.blogger.com/profile/05989277655934827117noreply@blogger.com