From cube-lovers-errors@oolong.camellia.org Thu Jun 5 16:09:12 1997 Return-Path: cube-lovers-errors@oolong.camellia.org Received: from oolong.camellia.org (localhost [127.0.0.1]) by oolong.camellia.org (8.6.12/8.6.12) with SMTP id QAA07093; Thu, 5 Jun 1997 16:09:12 -0400 Precedence: bulk Errors-To: cube-lovers-errors@oolong.camellia.org Date: Thu, 05 Jun 1997 08:41:16 -0400 (Eastern Daylight Time) From: Jerry Bryan Subject: Re: More on Korf's method In-reply-to: <199706050457.AAA19765@life.ai.mit.edu> To: cube-lovers@ai.mit.edu Message-id: MIME-version: 1.0 Content-type: TEXT/PLAIN; charset=US-ASCII X-X-Sender: jbryan@pstcc6.pstcc.cc.tn.us On Thu, 5 Jun 1997, michael reid wrote: > this seems to be the most fundamental difference between kociemba's > algorithm and korf's method: kociemba is interested in sub-optimal > solutions (optimal solutions are ok, too), whereas korf has no interest > in sub-optimal solutions. Good explanation, but I guess I still am unclear on one point. It seems that Kociemba's algorithm finds sub-optimal solutions which are either very close to optimal (or may actually be optimal -- by accident, as it were), and finds them very quickly. It also seems that Kociemba's algorithm will eventually find optimal solutions if it runs long enough, but "long enough" may be a long time. I think that I am hearing that "long enough" means that phase1 has essentially subsumed phase2 to the point that phase2 contains no moves. Is this correct -- that is, does the Kociemba algorithm guarantee us an optimal solution only after the solution is derived entirely in phase1? If not, at what point does the algorithm itself guarantee an optimal solution? = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = = Robert G. Bryan (Jerry Bryan) jbryan@pstcc.cc.tn.us Pellissippi State (423) 539-7198 10915 Hardin Valley Road (423) 694-6435 (fax) P.O. Box 22990 Knoxville, TN 37933-0990