Programming Challenge Competition

[+Vorticon]

3 hours ago, Asmusr said:

Does anything in the rules prevent you from scanning the entire maze before making your move?

Nope. That's a valid strategy to determine at least where the exit is.

[+Vorticon]

Ok so we have 4 competitors so far, so we have a contest! The poll will close on July 16th to give more of an opportunity for others to mull over the problem and decide if they want to participate. 

The contest will start officially on July 17 and run for 2 weeks.

I'll be announcing a prize this weekend.

 

[+Vorticon]

8 hours ago, unhuman said:

I'm thinking it's probably more than that, but I haven't programmed anything in TI in a while...  But...  I'm just kind of thinking the exit detection is actually going to make it more a game of chance than deterministic...

Yes, on the first run. The second run * should * be exclusively deterministic with no exploration, which is where the challenge lies.

[SteveB]

10 hours ago, Vorticon said:

Nope. That's a valid strategy to determine at least where the exit is.

If I can scan the maze I can determine the optimal way to the exit, just like Google Maps. 

Having only a local view like a mouse would have in a real maze, this is the actual challenge for me. 

I would suggest to restrict the view on the maze to the eight adjacent fields of your current position at all times. 

[+Vorticon]

3 hours ago, SteveB said:

If I can scan the maze I can determine the optimal way to the exit, just like Google Maps. 

Having only a local view like a mouse would have in a real maze, this is the actual challenge for me. 

I would suggest to restrict the view on the maze to the eight adjacent fields of your current position at all times. 

Good point. One could however argue that you are scanning the maze either way, whether through space by space exploration or scanning row by row. That said, since there is a step limit of 350, scanning will avoid that issue all together.

OK I think we'll disallow scanning of maze and I'll update the rules in the first post.

[+FarmerPotato]

To clarify, even on the first run, the mouse must move 1 square at a time? Say it hits a dead end, and you want to begin again at the start. It must take all the steps back? No “jumping”

 

[+Vorticon]

4 minutes ago, FarmerPotato said:

To clarify, even on the first run, the mouse must move 1 square at a time? Say it hits a dead end, and you want to begin again at the start. It must take all the steps back? No “jumping”

 

Imagine you are a mouse. What would you do? Same thing ?

[+FarmerPotato]

Will there be cheese and cats?

[unhuman]

So is the 350 total steps or total amount of tracking that can be done?  AKA if I step 349 to a dead end, valid to backtrack 348 and find a new path?

 

I think we need to solidify the rules here a bit - because there seems to be a bit of ambiguity. 

 

Seems like fun though!

[+Vorticon]

4 hours ago, unhuman said:

So is the 350 total steps or total amount of tracking that can be done?  AKA if I step 349 to a dead end, valid to backtrack 348 and find a new path?

 

I think we need to solidify the rules here a bit - because there seems to be a bit of ambiguity. 

 

Seems like fun though!

 

A step is any mouse movement of 1 space regardless of whether it is backtracking or not. 

Edited July 5, 2021 by Vorticon

[+Vorticon]

Prize announcement in the first post.

[unhuman]

I think my algorithm is in my head! I hope I don't forget it....

[+OLD CS1]

On 6/30/2021 at 10:22 AM, Vorticon said:

Update: the prize will be a brand new Logitech G300s gaming mouse (no pun intended ?)

I dunno.  I put one of these outside of a maze and it never moved.

[unhuman]

In a modern language working this through...  working with very small maps...  can't imagine 350 being anywhere near enough.

 

I think I'm almost done...  just gotta make my mouse brain.

[unhuman]

I'm a bit confused about the talk of scanning.  If we are scanning the maze before moving, then it's not the mouse doing the work - the mouse is then informed by Google Maps.

Is that really going to be a valid way to operate?

[+Vorticon]

1 hour ago, unhuman said:

I'm a bit confused about the talk of scanning.  If we are scanning the maze before moving, then it's not the mouse doing the work - the mouse is then informed by Google Maps.

Is that really going to be a valid way to operate?

Actually we ended up deciding against scanning the maze. The mouse will have to do all the work.

[+FarmerPotato]

I’ve got my mouse to explore the whole maze in pass 1, debugged.  and it knows how to repeat the direct course to the exit. But it can’t optimize around loops.. yet.  
 

Edited July 13, 2021 by FarmerPotato

[+Vorticon]

18 hours ago, FarmerPotato said:

I’ve got my mouse to explore the whole maze in pass 1, debugged.  and it knows how to choose a direct course to the exit. But it can’t optimize around loops.. yet.  
 

Excellent! We'll see how it will perform against the competition maze ? 

3 more days until we formally launch the contest.

[unhuman]

Will you / can you provide a sample maze?  Going to make assumptions about the dimensions otherwise? 

Really the only thing outstanding are the edges of the maze going to be 1/24, 1/32?

[+Vorticon]

6 hours ago, unhuman said:

Will you / can you provide a sample maze?  Going to make assumptions about the dimensions otherwise? 

Really the only thing outstanding are the edges of the maze going to be 1/24, 1/32?

Dimensions should not be relevant as long as the starting point is adhered to since all move decisions are made from the mouse's field of view which is 1 space in the 4 cardinal directions. The maze is bound by the '*' character all around except the entrance and exit and it's the same character used for the inner walls as well. All I can say about the maze is that it won't be limited to the classic maze design and the only absolute constants are an entry and an exit. You're going to have to work hard for that Logitech mouse ?

[unhuman]

Oh wow.  That is interesting...  probably different than I was working with...

 

So tricky - once you stumble out of the maze, you're done...  Nothing else to be done...

Edited July 14, 2021 by unhuman

[+FarmerPotato]

18 hours ago, unhuman said:

 

So tricky - once you stumble out of the maze, you're done...  Nothing else to be done...

I was wondering about that. The rules indicate that finding the exit is like a mouse trap! you’re done. So to explore fully, avoid the edges until ready. 

@Vorticon said there are * all around the border except for the entrance and exit. The mouse starts at X=16 and Y=23. That implies that there is a space at X=16, Y=24? Coordinates start at 1 not 0?

 

if the mouse GCHARs another space on the edge , it knows it’s the exit and can keep exploring, then? 

 

[+Vorticon]

8 hours ago, FarmerPotato said:

I was wondering about that. The rules indicate that finding the exit is like a mouse trap! you’re done. So to explore fully, avoid the edges until ready. 

@Vorticon said there are * all around the border except for the entrance and exit. The mouse starts at X=16 and Y=23. That implies that there is a space at X=16, Y=24? Coordinates start at 1 not 0?

 

if the mouse GCHARs another space on the edge , it knows it’s the exit and can keep exploring, then? 

 

All part of the challenge. And no, there is no space at X=16 Y=24. The entrance to the maze is essentially a nook in the exterior wall and the first step can only be to X=16 Y=22. That will prevent exiting from the entrance. 

The mouse can GCHAR one space in any of the 4 directions. If you know where the exterior wall is, then detecting the exit is easy, but the former is definitely not. Again, try not to make too many assumptions about the maze size or shape.

 

 

[unhuman]

Wow, this is now more of a riddle than a maze.... 

[+Vorticon]

OK I have updated the first post with clarified rules, a flow chart of the program flow, and submission information. Please let me know if anything remains unclear.

The competition launches tomorrow July 17 at 00:00 GMT. Good luck.