Implementation of a futoshiki game solver, with some (hopefully good) constraint satisfaction heuristics.

pypy <mode> solve-a-shiki.py [p] [<experiment file>]

where <mode> is one of:

• a: No backtracking heuristics applied
• b: Use forward checking for faster cutoff, when backtracking
• c: in addition to forward checking, use a 'least remaining values' policy to choose the next branch target position

The optional p flag activates profiling mode; perf-eval dumps are generated (with a much higher overhead). Also, if no <experiment file> is provided, the standard input is used (for manual input).

From Wikipedia

| Futoshiki (不等式 futōshiki?), or More or Less, is a logic puzzle game from | Japan. Its name means "inequality". It is also spelled hutosiki (using | Kunrei-shiki romanization).

Example of a 5x5 Futoshiki puzzle...

...And it's solution