Multi agent collaboration using distributed value functions
Resumen:
In this paper we present the use of distributed value function techniques to reach collaboration in a multiagent system. We apply this method in two different simulation environments: a mobile robot planning/searching task and an intelligent traffic system in an urban environment. In the case of the intelligent traffic system, results show an improvement with respect to a standard fix-time controller and local adaptive controllers. Trajectories for optimal search in an obstacle environment are obtained in the mobile robot case. Some variations to the actual algorithm are pointed out to suit our cases. We conclude discussing our future work.
| 2000 | |
|
Reinforcement learning Distributed system Mobile robot Traffic control |
|
| Inglés | |
| Universidad de la República | |
| COLIBRI | |
| https://hdl.handle.net/20.500.12008/20811 | |
| Acceso abierto |
| Sumario: | In this paper we present the use of distributed value function techniques to reach collaboration in a multiagent system. We apply this method in two different simulation environments: a mobile robot planning/searching task and an intelligent traffic system in an urban environment. In the case of the intelligent traffic system, results show an improvement with respect to a standard fix-time controller and local adaptive controllers. Trajectories for optimal search in an obstacle environment are obtained in the mobile robot case. Some variations to the actual algorithm are pointed out to suit our cases. We conclude discussing our future work. |
|---|
