Solving Futoshiki

[MrXuj]

I've been playing with my calc for a bit now, and I'm getting the hang of it, a bit more at least.

Several days ago I was shown a puzzle called a Futoshiki, and it's a wonderful thing, in my opinion.

However, as I am in possession of a calculator capable of executing programs, I thought why not make a program that solves these things?

I've been trying for the past two days, without any luck. I'm currently using matrices with 1 as > and 2 as <, and it's not going so well.

So I was wondering if anyone could help me out?

Or could this count as one of those brainteasers?

[Bhaliar]

So you want to make a program that will solve them or a program that will create them and you solve them? As a side note, it looks like sudoku, but with constraints.

[MrXuj]

The program I'm looking for is the solver itself.

And yes, it's somewhat like a sudoku with the 3x3 squares replaced with > and < as additional requirements besides the [1..2..n] numbers.

[Bhaliar]

Well my first idea would be to assign variables to each box. assign the given #'s then solve. A block of check like if the first row then 2nd then 3rd, then colmns, then diagonal. It would put numbers in an reassign them until the codes are succesful.

[calcdude84se]

I might be misunderstanding what you said, but that sounds a lot like brute-forcing.

It might be more efficient to solve it in a logical manner similar to how a person would do it, but the implementation would be complicated. (To get an idea, look at some of the better basic sudoku solvers)

[Bhaliar]

It does seem like brute forcing, but there would be a difference. When it puts the nubers in it would perform a check that looks for any similair nubers. Say you wanted to put a 5 in. It would check diagonally, then vertically for a 5.

[kinkoa]

i have never heard of futoshiki so good luck i think it looks interesting tho