Programming Puzzles Group
When
Occasional Mondays 7:00pm - 9:00pm
What
We are a group of guys who like puzzles and use programming to understand them better.
Where
Maan Coffee
northwest corner of Gongti. Table by the north door.
Wechat: 360414628
Email
Ernieernie@beijinggoclub.com
Meeting #4 - polyomino further questions June 23 2014
- we have code for simple additions
- will look at the pentomino "combination table"
Meeting #3 - Pentominos Continued - June 16, 2014
- 3 people
- A list of properties of Pentominos. For any grouping, could we come up with a property dividing them that way? Yes, at least a horrible, explicit way. But is there an interesting property for each one?
The things we'd like to discover about pentominos
- Formula for how many polyominos of N squares are there?
- Algebra for sets of polys
- Combine/separate shapes into polys of size N (some shapes are splittable in different ways)
- are there things about group theory we can apply? additive identity, multiplicative identity, elementary units?
general formalus for certain types of combinations
I + I_n for example is pretty easy, as is N+I_n. But varying both isn't easy, and as they bend on themselves it gets harder too. But it's a promising avenue - for a given shape it is easy to calculate what the formula for combinations with I_n is. How broadly can this be done?
Meeting #2 - pentominos
- 2 people.
Properties of Pentominos
- Groupings:
- edges
- perimeter
- inner corners
- outer corners
- longest edge
- covering square size
- "missing" squares to complete a rectangle
- etc.
- wikipedia pentominos
- naming them: we use F,I,L,N,P,T,U,V,W,X,Y,Z
- question: for every grouping, does there exist a property which produces it?
- counting pentominos - it's surprisingly unclear how many there are with N squares.
- multiplying (adding) pentominos. How many free combinations of X&I are there
- Families / relatedness graphs - grouping pentominos based on how many single-square moves are necessary to convert one to the other. For example, I=>P = 2, I=>F,L=1 etc. There doesn't seem to be any sense to it.
Meeting #1 - 2048
- We discussed the game 2048. 4 people.
- Maximum score reachable?
- "pure" version - only 2s, no random placement
- "evil" version
- developing an AI - strategies:
-"entropy" - minimize gradient of differences between adjacent squares
-"survival" - with enough depth to the search, just surviving should be the prime board evaluation function