From news-rocq!jussieu.fr!enst!freenix!sunqbc.risq.qc.ca!news.maxwell.syr.edu!feed1.news.rcn.net!rcn!not-for-mail Fri Nov 5 17:46:48 1999 Article: 10943 of rec.games.corewar Path: news-rocq!jussieu.fr!enst!freenix!sunqbc.risq.qc.ca!news.maxwell.syr.edu!feed1.news.rcn.net!rcn!not-for-mail From: Hillis Newsgroups: rec.games.corewar Subject: Re: GAs (was Re: IMT #2 Results) Date: Thu, 04 Nov 1999 22:55:10 -0500 Lines: 45 Message-ID: <3822551E.FAC7F371@erols.com> References: <381FA8F6.265C2B49@erols.com> <382020e1.1384736@NEWS.FREE4ALL.CO.UK> <3820FAF7.D7904742@erols.com> <38219064.19574935@NEWS.FREE4ALL.CO.UK> Reply-To: dbhillis@erols.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Trace: 40/NHTd8M78+4z3sSZJIoTkBjQDT9Q9UamS9lk6RQno= X-Complaints-To: abuse@rcn.com NNTP-Posting-Date: 5 Nov 1999 03:55:23 GMT X-Accept-Language: en X-Mailer: Mozilla 4.5 [en] (Win95; I) Xref: news-rocq rec.games.corewar:10943 Bjoern wrote: > > So do your warriors only mutate, or do they also 'mate' (by > crossover?)? > I use crossover too - I think it helps. > ... > But right now this gives me a thought - I benchmarked my warriors > against a fixed set of warriors, so perhaps a quick convergence was > inevitable... When I've scored against a fixed set of warriors, the population has always comes to a quick, djn-streaming convergence. I checked my log file and I now estimate that the program ran for about 2 weeks continuously. Here are 2 other warriors that were on the hill at the end of the run. ;redcode-b verbose ;assert 1 ;name h1_3.red ;author Dave Hillis ;strat - Created with a GA. ; Ilmari's Mini-Tournament #2. spl.i # -38, > 18 mov.i } -9, { 0 mov.i > -31, { -2 djn.i $ -2, $ -3 end 0 ;redcode-b verbose ;assert 1 ;name h1_38.red ;author Dave Hillis ;strat - Created with a GA. ; Ilmari's Mini-Tournament #2. mov.i { 34, > 29 mov.i { 9, } -1 mov.i > 21, } -1 djn.i $ -2, } -3 end 0 Dave