This thesis presents the development of a chess engine for the variant 5D Chess with Multiverse Time Travel, which extends conventional chess with the ability to move pieces through time and across multiple timelines. These characteristics substantially alter the structure of the game tree, making the direct transfer of existing chess-engine approaches non-trivial. The aim of the thesis is to investigate the applicability and adaptation of established methods used in conventional chess engines. The implemented engine employs bitboard-based state representations, move ordering, a position evaluation function, and game-tree search based on negamax with alpha-beta pruning. Additional search optimizations include quiescence search, principal variation search, and move-ordering heuristics. Due to the specific characteristics of 5D Chess, additional approaches were developed for turn generation, search-depth limitation, and handling of time-travel moves. The effectiveness of the applied approaches is evaluated by comparing different configurations of the developed engine.
|
