TY - GEN
T1 - Centralised and Decentralised Control of Video Game Agents
AU - Robinson, Samuel
AU - Barnes, Chloe M
AU - Lewis, Peter R.
PY - 2022/1
Y1 - 2022/1
N2 - In this paper, the game of partially observable Ms. Pacman is used as a sandbox to evaluate Artificial Neural Networks (ANNs) that control multiple opponents (i.e. the ghosts). Comparisons between one central ANN that controls all ghosts, and multiple distinct ANNs, each controlling one ghost, are made. The NEAT algorithm is employed to evolve the ANNs. We find that chasing Ms. Pacman and exploring the map are both harder behaviours to learn for a centralised controller than for decentralised control. Further, both centralised and decentralised approaches produce vastly different behaviours for exploring the map. Novel techniques for comparing networks are also explored.
AB - In this paper, the game of partially observable Ms. Pacman is used as a sandbox to evaluate Artificial Neural Networks (ANNs) that control multiple opponents (i.e. the ghosts). Comparisons between one central ANN that controls all ghosts, and multiple distinct ANNs, each controlling one ghost, are made. The NEAT algorithm is employed to evolve the ANNs. We find that chasing Ms. Pacman and exploring the map are both harder behaviours to learn for a centralised controller than for decentralised control. Further, both centralised and decentralised approaches produce vastly different behaviours for exploring the map. Novel techniques for comparing networks are also explored.
UR - https://link.springer.com/chapter/10.1007/978-3-030-87094-2_10
U2 - 10.1007/978-3-030-87094-2_10
DO - 10.1007/978-3-030-87094-2_10
M3 - Conference publication
SN - 9783030870935
SN - 9783030870942
T3 - Advances in Intelligent Systems and Computing
SP - 108
EP - 120
BT - Advances in Computational Intelligence Systems
PB - Springer
T2 - 20th UK Workshop on Computational Intelligence
Y2 - 8 September 2021 through 10 September 2021
ER -