Login | Register

RoboZZle Forums

Forums / Puzzles / Gm 101

Thread for puzzle Gm 101. Hints are OK, but please no full solutions.

 

Log in to comment.

http://www.salomonshoesoutlet.us.com
http://www.coachfactory-outletstore.us.com
http://www.uggsoutletboots.com.co
http://www.moncler-outletonline.us
http://www.pandora-jewelrycanada.ca
http://www.moncleroutlets.us.org
http://www.coachfactoryoutletsshop.us.com
http://www.pradaoutlets.us.com
http://www.airjordans.us.com
http://www.nike-shoes.in.net
http://www.newbalance-outlet.us.com
http://www.timberlandboots.in.net
http://www.poloralphlaurenoutletpolo.us.com
http://www.uggs-outlets.com
http://www.nikeairmax1.us
http://www.salvatoreferragamoshoes.com
http://www.canadagooseoutlet-stores.us.com
http://www.moncleroutletshop.us.com
http://www.nikeshox-shoes.com
http://www.ultraboost.us
http://www.nikeoutletshoes.us.com
http://www.adidascampusshoes.us
http://www.coach-factory-outlet.us.org
http://www.philipppleinshirt.com
http://www.nikeairmax95.com
http://www.suprashoes.name
http://www.katespadehandbagsclearance.us.com
http://www.converseshoes-outlet.us.com
http://www.toryburch-outletstores.us.com
http://www.louboutinshoes.in.net
http://www.michael-korshandbags.us.org
http://www.nike-basketballshoes.com
http://www.longchamps.us.com
http://www.adidas-yeezyshoes.us.com
http://www.nike-huarache.in.net
http://www.uggsoutletstores.us
http://www.jewelleryscharm.com
http://www.pandora-outlet.us.com
http://www.texasrangersjersey.us
http://www.atlantabravsjerseys.us
http://www.outlet-canadagoose.us
http://store.christianlouboutinsale.us.org
http://www.airmax-97.us.com
http://www.2017nikeairmax.us
http://www.guccioutlets.us.org 
http://www.coachcanadaoutlet.ca
http://www.nike-dunks.com
http://www.rolex-replicawatches.com.co
http://www.coachoutletstore-online.us.com
http://www.air-max90.org.uk
http://www.canadagooseoutletonline.us
http://www.katespade-outlet.us
http://www.cheapjordans.com.co
http://www.coachfactoryoutletonline.in.net
http://www.adidas-stansmith.com
http://www.michaelkors-outlet.in.net
http://www.losangelesdodgers-jerseys.us
http://www.pandorajewelryoutlet.in.net
http://www.pandora.us.com
http://www.pandoracharms.us.com
http://www.drmartens.us.com
http://www.michaelkorsoutletclearance.us.org
http://www.louboutinshoesuk.org.uk
http://www.coachfactoryoutletonlinesale.us.com
http://www.okaleysunglasseses.us.com
http://www.nike-store.us.com
http://www.pandoracharms.name
http://www.pandorajewelrypandora.us.com
http://www.longchamp-handbagsoutlet.us.com
http://www.fitflops-sale-clearance.us.com
http://www.timberlandoutlet.com.co
http://www.uggbootsonlines.us.com
http://www.toryburchoutlet-online.in.net
http://www.canadagoosejacketss.us.com
http://www.nikehuarache-shoes.com
http://www.toryburch-outletonline.in.net
http://www.cheapraybans.com.co
http://www.outletpandora.us.com
http://www.reebokshoes.us.org
http://www.cheapoakleysunglasses.com.co
http://www.coachfactory-outlet.us.com
http://www.adidas-shoes.us.com
http://www.christianlouboutinshoesuk.org.uk
http://www.kyrie4-shoes.com
http://www.charmspandora.us.com
http://www.outletuggs.us
http://www.northfacejacketss.us.com
http://www.birkenstocksandalsuk.co.uk
http://www.ugg-boots.us.org
http://www.suprashoes.us.org
http://www.raybansunglasses.name
http://www.underarmourshoesoutlet.us.com
http://www.birkenstockoutletus.us.com
http://www.outletcoachoutlet.us.com
http://www.coachfactoryoutlet-stores.us.com
http://www.birkenstocks.us
http://www.adidas-nmd.us.com
http://www.canadagooseoutletcanadagoose.us.com
http://www.uggsoutletofficials.us.com
http://www.adidasoriginals.in.net
clb20181108

clibin009, 10 months ago.

20180723lck
http://www.ysloutlet.us.com
http://www.lightupshoes.us.com
http://www.lakersjersey.us.com
http://www.michaelkorsoutletnew.us.com
http://www.chicago-blackhawks.us.com
http://www.coachoutletonlineshops.us.com
http://www.poloralphlaurencheap.us.com
http://www.jackwolfskin.us.com
http://www.jimmychoo-sunglasses.us
http://www.ferragamooutletstore.us.com
http://www.katespadeoutletkate.us.com
http://www.swarovskioutletonline.us.com
http://www.nikelunarglide.org
http://www.christianlouboutinoutletin.us.com
http://www.michaelkorsoutletmks.us.com
http://www.hermesoutletonline.us.com
http://www.pandoraoutletjewelry.us.com
http://www.katespadeoutletofficials.us.com
http://bulls.nbajerseysstore.us.com
http://www.nikecortezshoes.org
http://www.alifeclothing.us.com
http://www.chelseajerseys.us
http://www.polooutletfactorystore.us.com
http://www.katespadeoutletofficial.us.org
http://www.jeanstruereligion.in.net
http://www.coachoutletclearance.us.com
http://www.swarovskioutletinc.us.com
http://www.uggoutletstoresofficial.us.com
http://www.belstaffjackets.us
http://www.salvatore-ferragamo.in.net
http://www.canadagoosejacketssales.us.com
http://grizzlies.nbajerseysstore.us.com
http://www.guccioutletstores.us.com
http://www.uggoutlet-onlineclearance.us.com
http://www.airmax97.net
http://www.canadagooseoutlet-store.name
http://www.adidascrazy.us.com
http://www.canadagoose-jacketsuk.co.uk
http://www.michaelkorsukhandbags.co.uk
http://www.oakleysunglassescase.us.org
http://www.raybansunglassessale.net.co
http://www.poloralphlaurenwholesale.us.com
http://www.coachoutletonlineco.us.com
http://www.adidas-uk.me.uk
http://www.michaelkorsoutletca.us.com
http://www.cheapjordansretro.in.net
http://www.moncler-jackets.org.uk
http://www.fendihandbags.us.com
http://www.manchesterunitedjersey.us
http://www.armaniexchange.us.org
http://www.adidasnmduk.org.uk
http://www.canadagoosejacketsgoose.us.com
http://www.coachfactoryoutletclearance.us.com
http://www.gstarjeans.us
http://www.jerseyfromchina.us.com
http://www.airmax90nike.org.uk
http://www.christianlouboutinoutletcl.us.com
http://www.fitflopssale.in.net
http://www.suprashoes.us.com
http://www.asicsshoes.net
http://www.basketballshoes.us.com
http://www.kd10nike.us.com
http://www.lebronjamesshoess.us.com
http://www.canadagoosegoose.com.co
http://www.newbalanceshoes.in.net
http://www.christianlouboutinoutletsale.us.com
http://www.nikeairhuaraches.co.uk
http://knicks.nbajerseysstore.us.com
http://www.kobebryantshoes.in.net
http://www.canadagooseoutletonlinesale.us.com
http://www.cheap-jordanshoes.us.com
http://www.coachoutletstore.com.co
http://www.nikeoutletonline.us.org
http://www.coachoutletonlinefactorystore.us.com
http://www.toryburchstoreoutlet.us.com
http://www.long-champsolde.fr
http://www.raybansunglassesoutlets.com.co
http://www.coachoutletonline.us.org
http://www.wellensteyn.us.org
http://www.jordanshoesstore.us.com
http://www.futbol-baratas.com
http://www.airjordanretro.com.co
http://www.dsquared2jeans.us.com
http://www.pradashoesofficial.us
http://pacers.nbajerseysstore.us.com
http://www.coach-handbags.us.com
http://jazz.nbajerseysstore.us.com
http://www.mulberry-handbagsoutlet.org.uk
http://www.canadagoosejacketsstore.us.com
http://www.jeepshoes.us
http://www.salomonshoess.us.com
http://mavericks.nbajerseysstore.us.com
http://www.uggoutlets.ca
http://www.outlettruereligion.in.net
http://www.clarks.us.org
http://www.oakleysunglassesorg.us.org
http://www.ferragamo-outlets.us.com
http://www.brequetwathes.us
http://www.michaelkorsoutletcheapsale.us.com
http://www.michaelkorsoutlets.me.uk
20180723lck

liuchunkai, 13 months ago.

easy cheezy; hint: 1 conditional function

morgsis's comment after solving the puzzle, 50 months ago.

1287 reuqires some initial idea so I still didn't solve it:)

There are not much puzzles that I solved using bruteforcer, much more often I just look for the shortest solution.

BTW once I found length 9 solution for puzzle 14 and was surprised how did bot walk and how is it possible to find such solution by mind. (I could find only length 10 solution with much simpler route.) Slightly later I similarly found length 9 solution for puzzle 133 and I had strong déjà vu feeling, looking how did bot walk. I checked that I didn't bruteforced this puzzle earlier (I save input and output files) and was really puzzled by this déjà vu feeling. Later I accidentally opened puzzle 14 and noticed that it is identical with 133. It was relief for me:)

IMHO sharing source code of such took to all is bad idea. You will give some advantage to people who can compile code, but possibly can't think:) It is trivial to write naive solver, so everyone who needs should write it. Obviously if someone can solve any hard puzzle (s)he have good programming potential (possibly just need to learn coding), 'cause robozzle is really esoteric programming language. And it is more challenging to implement anything in robozzle language than in some other.

smylic, 94 months ago.

I'm still only in the 8-9 range, 10 if I wait a few days. Most of the optimisations I use are only to prune the search tree slightly. I'm using JIT compilation of the robozzle programs so that I can run them faster, but that's only a constant factor improvement.

The only current avenue of improvement I have is finding a better way of partitioning the search space to make use of multiple cores, my current method is inefficient, but again that's only a constant factor improvement. I'd love to know what your idea is Smylic.

I've also solved a couple of Axorion's puzzles, where I had most of the commands, but needed a couple more "random" commands to finish it (1287 and 1978). I did this by hardcoding many of the commands, which greatly reduced the search space. I did this by writing a custom version of the search tree generator for each puzzle, so it's not really automated yet.

I've considered putting my source code up on github or something, not sure how the community would feel about that. Its use would be limited to people who know how to compile. I regularly use it to solve puzzles that I consider "not fun", such as random painting.

keba, 94 months ago.

Thanks for the response, I'm very curious how the idea turns out. Even getting to 11 or 12 would be a major accomplishment for a solver I'd think (probably a 50x speedup).

At some point I should really work on my solver more, I'm sure I could get it to 8-9 length without too much difficultly (technically it could already solve an 8 if I let it run a day or two, but I'm too impatient to let it run that long).

But, for now every time I have free time I always choose to solve more puzzles, since I'm so far behind after taking ~1 year off twice. I kind of alternate between blitzing through as many long (20+ move) puzzles as I can and solving and optimizing short ones.

azure, 94 months ago.

Oh, I have good idea for solver for about 1.5 years, but still didn't realize it. My actual solver is much slower, it finds length 8 solutions for puzzles with three colors (and up to 4 functions, or less functions, but with paint) or length 9 solutions for puzzles with two colors.

I don't know why I said about length 17–19, it seems I meant very special case. Even possibility of solving all puzzles with length 12 solution and two colors/two functions should be a great deal. But I hope that length 14–15 solution is bruteforcable in some rather frequent case of two colors/two functions puzzle. I promise to implement my idea next week and check my supposition :)

smylic, 94 months ago.

Is there any forum thread or puzzle thread about solvers? I realize for some it is "cheating", but I'm extremely interested in the topic. A good solver is a major programming challenge and most us of like robozzle because of its similarity to programming challenges.

Not sure smylic or keba will read this, but I can't even fathom a program that could solve more than maybe length 12 on a two color puzzle. The public solver (which is admittedly quite slow) could barely do length 7 when I tried it 2-3 years ago and I don't think it's changed.

I know optimizations can eliminate a huge number of combinations, but presumably most optimizations don't get exponentially better with more length. At 2 colors, 2 functions, each additional command is 15 times slower.

Assuming you could make a program a million times faster than public one that would still barely be depth 12. If a solver could really do even a length 15 than it must be able to solve a length 12 nearly instantly which doesn't seem remotely possible.

azure, 94 months ago.

You probably won't read this reply, but they're talking about a programs they wrote that bruteforce the puzzle solutions (test every combination of commands). This is possible for puzzles with few commands or simple puzzles with zero or very few conditionals, but for complex puzzles, our computers are not powerful enough to test the many combinations in a reasonable amount of time.
Sometimes I wonder how many solvers use bruteforcers and how often.
I shamelessly will admit here that I didn't bruteforce this puzzle and have never bruteforced any puzzle using a program! (I have manually tested possible solutions until a real solution was found)
I found a 3, 4 solution to this puzzle that was not intuitive but wasn't too difficult to find; a 2, 5 solution is not obvious to me.

sticky, 94 months ago.

I feel like such a noob when I look at these three's chatter... anyone feel the same?

gm101, 94 months ago.

I found bug. And it was in my input file with puzzle description... :D

@Keba: I think length 17-19 may be bruteforcabele for some puzzles with only two colors, only two functions and without paint. I'm short of time now so I didn't try to implement my main optimization idea which should be awesome for many puzzles. When I implement it I'll show you some puzzle with long solution that I can brute force:)

Smylic, 108 months ago.

There's multiple 3+4 and 2+5 solutions (total 22), although really only 3 main variants. And all hard to find without brute forcer, kudos if you do :-)

Also, I think your estimate of being able to brute force up to length 20 is far too high. I can handle length 10 without painting, maybe a couple more if I am very patient, but the exponential search space means length 20 is impossible even if I am missing substantial optimisations.

keba, 108 months ago.

I have verified my solution and it is correct. Keep on debugging :)

edsaw, 108 months ago.

edsaw, do you really have correct solution with length 7?
After finding length 8 solution I tried to optimize it, but couldn't. Then I used my bruteforcer. It didn't find such solution... I switched off all my optimizations that speed up finding solution to make it try every imaginable sequence of 3+4, 2+5 and 1+6 commands. Senselessly... I reread code and didn't find any bug... So at the moment I don't believe that there is length 7 solution;)

Smylic, 108 months ago.

Previous page (1 / 1) Next page

©2009 Igor Ostrovsky and other contributors. Terms of use