Kirjoitettu: 4. tammikuuta 2006 klo 02:27
* The search algorithm is a min-max search with alpha-beta pruning
* HiddenLynx uses iterative deepening to maintain nearly constant thinking time for each move. The iterative deepening is enhanced with the principal variation line (i.e. the previously best branch is searched first when the next deepest level is searched)
* To further speed up the search, HiddenLynx uses killer heuristics (i.e. it considers the moves of the strongest pieces first)
* The evaluation of the board is based on the "mass" of the pieces and an additional position map, which ranks strong positions for each piece type with a higher value.
* On a computer with a 3.2 GHz CPU, Hiddenlynx searches roughly 7-11 plies in 10 seconds.
Imuroi tiedosto
