Jigsaw Hybrids

[gabriele]

I'm trying to make a valid grid of a jigsaw+variant.
In spite of thousands of tentatives with each variant, i couldn't make one valid.
why? why the computer can not realize a valid jigsaw+untouch grid? It should be easy!
now, I hope to have found the answer.
if you don't take into account the in&out technique
https://www.argio-logic.net/forum/viewtopic.php?f=68&t=1987,
the generation process will have to process millions of combination before finding a valid one.
now I'll try to translate the in&out tecnique into a routine.

Sto tentando di produrre una griglia valida di jigsaw + variante.
nonostante migliaia di tentativi con ogni variante il programma non riesce a generare una griglia valida.
perchè?
come mai il programma non riesce a generare una griglia valida? dovrebbe essere facile!
penso di aver trovato la risposta.
se non si tiene in considerazione la tecnica dell in&out
https://www.argio-logic.net/forum/viewtopic.php?f=68&t=1987,
il programma dovrà generare milioni di griglie prima di trovarne una valida.
ora proverò a trasformare la tecnica in algoritmo.

[gabriele]

4 days to find the routine for the in&out tecnique,
I was surprised by the number of in & out area that you can find in a jigasaw grid , more than 30!
But all this work has not brought to any solution.
the program is not able to generate the 81 digits that respect the rules of 2even 2odd on the following grid.

does someone succeed in producing a valid grid respecting the rules of 2 even 2 odd ?



4 giorni per produrre la routine della tecnica in e out.
sono rimasto sorpreso dalla quantità di zone interne-esterne che si possono trovare in una griglia, più di 30.
ma tutto questo lavoro non è servito a niente!
il programma non riesce a trovare gli 81 numeri che soddisfino anche le regole del 2 even 2 out.
qualcuno ci riesce?

[gabriele]

You can draw an imaginary line on the border of any row or column. The cells which goes OUT of the line and the cells which comes IN must contain the same numbers.
In any jigsaw grid, an imaginary line can be drawn at any row/column... but it is effective at certain places only...

Can somebody tell me why it is effective at certain places only?

ore tell me where I can find more information?

[Fred76]

You can draw an imaginary line on the border of any row or column. The cells which goes OUT of the line and the cells which comes IN must contain the same numbers.
In any jigsaw grid, an imaginary line can be drawn at any row/column... but it is effective at certain places only...

Can somebody tell me why it is effective at certain places only?

ore tell me where I can find more information?

To my knowledge, it's available for any row/column, but perhaps you don't have to use all these informations to solve grids. The technic is called LoL (Law of Leftovers):
http://www.sudopedia.org/index.php?title=LoL

Gabriele, it's more difficult to create jigsaw, and especially jigsaw+variant, than classical grid. constraints are big on jigsaw. When you create grids by hands, you see very well what I mean...

For example, there are some region configurations that can't have any solutions. Look at this grid:


That grid has no solution, simply because of disposition and region's shapes.

Fred

[gabriele]

Thanks Fred,
I knew that jigsaw grid are limited
https://www.argio-logic.net/forum/viewtopic.php?f=43&t=1352
but I'm very surprised of
1 i'm (my vb project) not able to produce any jigsaw grid + variant (that adds a costraint)
2 the several in & out areas that you can find in a grid. I'm not sure if they are valid areas, tomorrow I will post some to submit to your opinion.

[gabriele]

Why is quite impossible to generate a valid grid of jigsaw+ constraint variant?
I am not entirely sure of my affirmations, also because my interpretation of the in & out rule could be wrong.
I made a test starting with a grid with just few numbers entered:
the in&out tecnique at work

[gabriele]

Now i run a brute force,

after 12 hours of work the answer is

no valid grid found that satisfy the rules of 2 even 2 odd

[Bely]

But since i like jigsaw, please Boss, keep going on with the generation of these hybrids... for example with diagonal would be surely possible (I have seen it somewhere, although I don't remember where )...would be nice with nonconsecutive or antiknight, but I am not able to figure out if it is ok or not

[Gotroch]

I think that the reason why you had a little success in generating valid jigsaw+2odd2even grid is because you didn't pay enough attention to odd/even part of the puzzle.

It looks entering two rows of "random" digits could be a good way to choose given digits for that puzzle, but you can easily run into situation, that puzzle doesn't have any solutions.

When I tried to create jigsaw+2odd2even sudoku I used that way:
First I need jigsaw grid... I choose the one you tried.
Then I start marking each square in the puzzle with two colours, one for odd numbers and one for even numbers to create grid which satisfies 2odd2even rules - see picture 1. (Each row, column and region has 5 blue(odd) squares and 4 pink(even) squares
and there aren't 3 odd or 3 even digits in a line in the grid - It's easy to choose the squares manually and I guess it shouldn't be that difficult let your generator to do it)
With such prepared grid you can start filling it with numbers and create valid puzzle - picture 2

[gabriele]

Gotroch

the great

[gabriele]

It's easy to choose the squares manually and I guess it shouldn't be that difficult let your generator to do it)

Yakub,you remember me when, still young, I studied mathematical analysis at the university.

There was anything that frightened me more of when, in the book of the teacher, I found writing: "there is no need to demonstrate that....."

Usually that sentence was prelude to days of study on the enunciate.

I don't tell you how difficult was to choose the squares manually in just 1 grid, something left to the case, so how do you think that I can translate in a program code something that I'm not able to figure in my mind.

But sooner or later i will.

[gabriele]

ok,
I have succeeded in producing the program that identify valid sequences of even-odd cells.
but here is the surprise: not always the sequences are valid even if they respect the odd -even position.
can you see why the following grid is not valid?
hint: look at the even numbers
moral: still I am a lot away from producing a valid grid

[gabriele]

ok,
I have succeeded in producing the program that identify valid sequences of even-odd cells.
but here is the surprise: not always the sequences are valid even if they respect the odd -even position.
can you see why the following grid is not valid?
hint: look at the even numbers
moral: still I am a lot away from producing a valid grid

this is why the sequence is NOT valid
we start with "2", but any other even number will bring to the same result








the "2" is not possible in row

[gabriele]

as promised
"sooner or later"


I would like to have your opinion about the difficulty;
Is the "in & out" technique (law of leftovers) necessary?

[genovese volante]

I did it.....
I think LoL is not necessary......
But there's a simple mind step that
Camillo could considere a "Trial&error".....

[gabriele]

Jigsaw Hibrids are now
available daily