DocumentCode
445834
Title
An evolved Seega player capable of strong novice-level play
Author
Abdelbar, Ashraf M. ; Soliman, Ossama ; Kinawy, Shefif ; Sayed, Hisham
Author_Institution
Dept. of Comput. Sci., American Univ. in Cairo, Egypt
Volume
1
fYear
2005
fDate
31 July-4 Aug. 2005
Firstpage
332
Abstract
Seega is an ancient Egyptian two-stage board game that, in certain aspects, is more difficult than chess. The two-player game is most commonly played on a 7 × 7 board, but is also sometimes played on a 5 × 5 or 9 × 9 board. In the first and more difficult stage of the game, players take turns placing one piece each on the board until the board contains only one empty cell. In the second stage players take turns moving pieces of their color; a piece that becomes surrounded by pieces of the opposite color is captured and removed from the board. A player wins when he has captured all, or all except one, of his opponent´s pieces. We present results on the evolution of a Seega player that plays at the level of college students, who have had only a few days´ familiarity with the game. The player is based on minimaxing with 12 hand-coded features, and feature weights that are evolved by a co-evolutionary system comprising a particle swarm optimizer (PSO) and an evolutionary algorithm (EA). Separate White and Black populations of feature weight vectors are maintained; each population is evaluated based on play against players from the top-, middle-, and bottom-third of the other population.
Keywords
evolutionary computation; particle swarm optimisation; ancient Egyptian two-stage board game; coevolutionary system; evolutionary algorithm; evolved Seega player; particle swarm optimizer; strong novice-level play; Aggregates; Computer science; Dogs; Educational institutions; Evolutionary computation; Heart; Minimax techniques; Particle swarm optimization; Vectors;
fLanguage
English
Publisher
ieee
Conference_Titel
Neural Networks, 2005. IJCNN '05. Proceedings. 2005 IEEE International Joint Conference on
Print_ISBN
0-7803-9048-2
Type
conf
DOI
10.1109/IJCNN.2005.1555852
Filename
1555852
Link To Document