A graphics interface and an evaluation of parallel alpha-beta search algorithms for a mechanical bridge player

As one of a series of Master's theses that will hopefully lead to the development of a world class contract bridge playing program, this work focuses on two areas: the design and implementation of the graphics interface CBRIDGE and an introductory exploration of the potential benefits of three parallel alpha-beta search algorithms. First, CBRIDGE was designed and implemented in order to provide an environment in which human and machine players could interact. This interface will facilitate the testing of machine players and eventually enable machine players to be entered into tournaments. Second, we conducted experiments with three parallel alpha-beta search algorithms -- aspiration, principal variation alpha-beta and principal variation splitting -- that were developed with chess game trees in mind to see whether the reported benefits in speedup applied to the search of bridge game trees. We anticipated these results would indicate what magnitude of improvements would be possible by parallelizing the alpha-beta search currently used in our double-dummy machine player.