Net_Structure

Tips

参考文献

1. we can learn a fully centralised stateaction value function Q_tot and then use it to guide the optimisation of decentralised policies in an actor-critic framework
2. QMIX consists of agent networks representing each Qa,
   and a mixing network that combines them into Q_tot, not
   as a simple sum as in VDN, but in a complex non-linear way that ensures consistency between the centralised and decentralised policies
3. non-linear mixing of agent Q-values in
   order to achieve consistent performance across tasks.
4. cooperative setting

constraint

This function allows each agent to participate in a decentralised execution by choosing greedy actions with respect to its value function.