The neural-network algorithm is studied to train AI players for push-ball game. Two competitive teams can assign different number of players to investigate the cooperation of members. In this version, only the common score criterion and the trivial interactive model among team members are applied. The neural-network algorithm is referred to Peter Bone's MatLab code "Machine learning - balancing a stack of balls"
I-Lin Ho (2019). Machine learning - PushBall game (https://www.mathworks.com/matlabcentral/fileexchange/65350-machine-learning-pushball-game), MATLAB Central File Exchange. Retrieved .
modify the physical model
Inspired by: Machine learning - balancing a stack of balls