[reinforcement learning notes] common symbols in reinforcement learning

Symbol	Symbol interpretation
$\doteq$	Defining symbols
$\approx$	About equal to
$ϵ$	$ϵ$ The probability of random action in a greedy strategy
$\gamma$	Discount factor
$\lambda$	Decay rate in trace
$\leftarrow$	Assignment symbol
$s$, $s^{\prime }$	state $s$
$a$	action $a$
$r$	earnings $r$
$t$	Discrete time steps , Or time
$\pi$	Strategy （ Decision making rules ）
$\pi (s)$	according to Deterministic strategies $\pi$ In state $s$ Action selected when
$\pi (a|s)$	according to Random strategy $\pi$ In state $s$ Action selected when $a$ Probability
$A_t$	$t$ The action of the moment
$S_t$	$t$ The state of the moment , Usually by $S_{t-1}$ and $A_{t-1}$ Random decision
$R_t$	$t$ The benefits of the moment , Usually by $S_{t-1}$ and $A_{t-1}$ Random decision
$G_t$	$t$ The reward of the moment （ It's an expectation ）
$p(s^{\prime },r|s,a)$	From the State $s$ Take action $a$ Move to state $s^{\prime }$ And get the benefits $r$ Probability
$p(s^{\prime }|s,a)$	From the State $s$ Take action $a$ Move to state $s^{\prime }$ Probability
$r(s,a)$	From the State $s$ Take action $a$ The expectation of immediate benefits
$r(s,a,s^{\prime })$	From the State $s$ Take action $a$ Move to state $s^{\prime }$ The expectation of immediate benefits
$v_{\pi }(s)$	state $s$ In the strategy $\pi$ Under the value of （ Expected return ）
$v_{\ast }(s)$	state $s$ The value under the optimal strategy
$q_{\pi }(s,a)$	state $s$ In the strategy $\pi$ Take action $a$ The value of
$q_{\ast }(s,a)$	state $s$ Take action under the optimal strategy $a$ The value of
$V$, $V_t$	State value function
$Q$, $Q_t$	Action value function