Distributed Intelligent Learning and Decision Model Based on Logic Predictive Control
Algorithm 2
Model prediction action selection mechanism algorithm.
(1)
Initialization.
(2)
Repeat the following steps:
(a)
Observe the current environmental state s.
(b)
Select actions according to the improved action selection mechanism.
(c)
Observe and record the new state s′, individual action A, other agent actions, and instant reward; then, update the state prediction function and state transition probability function, and calculate the strategy change function of agent J.
(d)
Enter (e) if the target is reached; otherwise, return to (a).
(e)
Compute the Q function using the priority scan algorithm.
