From: Mark <mark.ericson@...>
Date: 2006-01-02T05:04:38+09:00
Subject: Re: [SOLUTION] Numeric Maze (#60)

Nevermind, a copy and paste error on my part!

On 1/1/06, Maurice Codik <maurice.codik@gmail.com> wrote:
> Mark: could you explain what you mean? these were outputs from that prog:
>
> solve(3,1233):
> [3, 6, 8, 16, 18, 36, 38, 76, 152, 154, 308, 616, 1232, 2464, 2466, 1233]
> solve(1,25):
> [1, 3, 6, 12, 24, 48, 50, 25]
>
> looks like it finds them just fine... or do u mean that these arent the
> optimal paths?
>
> James: Now that I think about it, your optimization and mine are almost
> equivalent, I think-- I'm fairly certain that the only way to get to a
> number that you've been to already is by using consecutive pairs of
> double/half.. since the add_two operation isn't invertable in this quiz.
>
> Maurice
>
> On 1/1/06, Mark <mark.ericson@gmail.com> wrote:
> >
> > Maurice,
> >
> > I just noticed that your solution can't find the path between two odd
> > numbers.
> >
> > - Mark
> >
> > On 1/1/06, Maurice Codik <maurice.codik@gmail.com> wrote:
> > > I guess we're allowed to submit solutions now... here's my first ever
> > ruby
> > > quiz solution (these quizzes are a great idea, btw):
> > >
> > > It's is an iterated-depth DFS w/ memoization written in a functional
> > style
> > > for fun.
> > >
> > > It exploits the fact that the optimal path through the maze will never
> > have
> > > consecutive double/halves, which reduces the avg branching factor of the
> > > search tree to ~2. Keeping track of this state makes the code a little
> > > uglier, but faster on the longer searches.
> > >
> > > Maurice
> >
> >
>
>