主催: The Japanese Society for Artificial Intelligence
会議名: 2013年度人工知能学会全国大会(第27回)
回次: 27
開催地: 富山県富山市 富山国際会議場
開催日: 2013/06/04 - 2013/06/07
The high space complexity and the high time complexity are still severe limitations when the game tree is searched until a very deep depth in computer gaming programs. In order to reduce these limitations, we aim to parallelize computer gaming programs by using the grid computing system which was developed by Professor I-Chen Wu. In this system, it exploits spare resources such as computing powers and memory storages of desktops or some personal computers for the applications requiring huge amount of computation. However, when the gaming programs are combined with the grid computing system, it is necessary to resume the gaming programs for each move during a game. To overcome the above problem, in this paper, we introduce some approaches to avoid the resumption of the running programs. According to the initial property of the computer gaming programs, we design two frameworks to combine the computer gaming programs with the grid computing system. The experiments show that its performance is improved 8.27 times on a Connect6 program and its winning rate is enhanced 42.8% on a Go program. The results indicate that our approach combined with the grid computing system is quite powerful for the computer gaming programs. Furthermore, these frameworks are easy for integration between the computer gaming programs and the grid computing system.