Boltzmann Machine Solution N-Person Games
Games with limited interaction are considered. A global behaviour of a team of players in the game is observed. The average total payoff received by the team of players is chosen as a criterion of the global behaviour. The game is implemented by a neural network of Boltzmann Machine type. The model is applied to solve an optimal task assignment problem in parallel computers. Results of experimental study of the model are presented.
Unable to display preview. Download preview PDF.
- Indurkhya B, Stone HS and Xi-Cheng L (1986) Optimal partitioning of randomly generated distributed programs. IEEE Trans. on Software Eng., SE-12(3):483–495Google Scholar
- Seredynski F (1991) Self-organization in homogeneous networks of learning automata. Proc. of the 6th Czechoslovak Annual AI Conference, Prague, June 25-27, 1991Google Scholar
- Seredynski F (1992) Mapping by migrating Boltzmann Machine. Proc. of Parallel Problem Solving From Nature Conference, Brussels, Sept. 28-30, 1992Google Scholar