Title :
Porting a computer GO program to the Windows CE platform
Author :
Cao, Huifang ; Liu, Zhiqing
Author_Institution :
BUPT-JD Res. Inst. of Comput. GO, Beijing Univ. of Posts & Telecommun., Beijing, China
Abstract :
Computer GO has seen a significant development since the UCT algorithm was used successfully in computer GO game tree search in 2006. Monte Carlo Game Search based on the UCT algorithm has since become the standard approach in all modern computer GO programs, in which computer GO problems are solved numerically as optimization problems of a large Markov decision process. It is challenging to port modern computer programs to an embedded system such as the Windows CE platform because extremely limited computing and memory resources are available to conduct numerical computation. This paper describes an approach consisting of the following three steps to port lingo, a modern computer GO program to the Windows CE platform so that lingo can run on mobile phones and other embedded systems: Firstly, expert knowledge is extensively used in the Windows CE based computer GO program in order to reduce its reliance on numerical computation; secondly, we reduce the search space of game trees to meet the strict memory requirements in the Windows CE platform; and lastly, a graphical user interface is ported to Windows CE.
Keywords :
Monte Carlo methods; artificial intelligence; computer games; embedded systems; graphical user interfaces; mobile computing; mobile handsets; operating systems (computers); optimisation; tree searching; Markov decision process; Monte Carlo game search; UCT algorithm; Windows CE platform; computer GO game tree search; computer GO program porting; embedded system; expert knowledge; graphical user interface; lingo; memory resource; mobile phones; numerical computation; optimization problems; Algorithm design and analysis; Computers; Embedded systems; Engines; Games; Markov processes; Monte Carlo methods;
Conference_Titel :
Control and Decision Conference (CCDC), 2011 Chinese
Conference_Location :
Mianyang
Print_ISBN :
978-1-4244-8737-0
DOI :
10.1109/CCDC.2011.5968636