Reinforcement Learning with Code 【Chapter 7. Temporal-Difference Learning】

news2024/11/14 5:49:49

Reinforcement Learning with Code

This note records how the author begin to learn RL. Both theoretical understanding and code practice are presented. Many material are referenced such as ZhaoShiyu’s Mathematical Foundation of Reinforcement Learning, .

文章目录

  • Reinforcement Learning with Code
    • Chapter 7. Temporal-Difference Learning
    • Reference

Chapter 7. Temporal-Difference Learning

Temporal-difference (TD) algorithms can be seen as special Robbins-Monro (RM) algorithm solving expectation form of Bellman or Bellman optimality equation.

7.1 TD leaning of state value

​ Recall the Bellman equation in section 2.2

v π ( s ) = ∑ a π ( a ∣ s ) ( ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v π ( s ′ ) ) (elementwise form) v π = r π + γ P π v π (matrix-vector form) v_\pi(s) = \sum_a \pi(a|s) \Big(\sum_rp(r|s,a)r + \gamma \sum_{s^\prime}p(s^\prime|s,a)v_\pi(s^\prime) \Big) \quad \text{(elementwise form)} \\ v_\pi = r_\pi + \gamma P_\pi v_\pi \quad \text{(matrix-vector form)} vπ(s)=aπ(as)(rp(rs,a)r+γsp(ss,a)vπ(s))(elementwise form)vπ=rπ+γPπvπ(matrix-vector form)

where

[ r π ] s ≜ ∑ a π ( a ∣ s ) ∑ r p ( r ∣ s , a ) r [ P π ] s , s ′ = ∑ a π ( a ∣ s ) ∑ s ′ p ( s ′ ∣ s , a ) [r_\pi]_s \triangleq \sum_a \pi(a|s) \sum_r p(r|s,a)r \qquad [P_\pi]_{s,s^\prime} = \sum_a \pi(a|s) \sum_{s^\prime} p(s^\prime|s,a) [rπ]saπ(as)rp(rs,a)r[Pπ]s,s=aπ(as)sp(ss,a)

​ Recall the definition of state value: state value is defined as the mean of all possible returns starting from a state, which is actually the expectation of return from a specific state. We can rewrite the above equation into

v π = E [ R + γ v π ] (matrix-vector form) v π ( s ) = E [ R + γ v π ( S ′ ) ∣ S = s ] , s ∈ S (elementwise form) \textcolor{red}{v_\pi = \mathbb{E}[R+\gamma v_\pi]} \quad \text{(matrix-vector form)} \\ \textcolor{red}{v_\pi(s) = \mathbb{E}[R+\gamma v_\pi(S^\prime)|S=s]}, \quad s\in\mathcal{S} \quad \text{(elementwise form)} vπ=E[R+γvπ](matrix-vector form)vπ(s)=E[R+γvπ(S)S=s],sS(elementwise form)

where S , S ′ S,S^\prime S,S and R R R are the random variables representing the current state, next state and immediate reward. This equation also called the Bellman expectation equation.

​ We can use Robbins-Monro algorithm introduced in chapter 6 to solve the Bellman expectation equation. Reformulate the problem that is to find the root v π ( s ) v_\pi(s) vπ(s) of equation g ( v π ( s ) ) = v π ( s ) − E [ R + γ v π ( S ′ ) ∣ S = s ] = 0 g(v_\pi(s))=v_\pi(s) - \mathbb{E}[R+\gamma v_\pi(S^\prime)|S=s]=0 g(vπ(s))=vπ(s)E[R+γvπ(S)S=s]=0, where S ′ S^\prime S is iid with s s s. We can only get the measurement with noise ​

g ~ ( v π ( s ) , η ) = v π ( s ) − ( r + γ v π ( s ′ ) ) = v π ( s ) − E [ R + γ v π ( S ′ ) ∣ S = s ] ⏟ g ( v π ( s ) ) + ( E [ R + γ v π ( S ′ ) ∣ S = s ] − [ r + γ v π ( s ′ ) ] ) ⏟ η \begin{aligned} \tilde{g}(v_\pi(s),\eta) & = v_\pi(s) - (r+\gamma v_\pi(s^\prime)) \\ & = \underbrace{v_\pi(s) - \mathbb{E}[R+\gamma v_\pi(S^\prime)|S=s]}_{g(v_\pi(s))} + \underbrace{\Big( \mathbb{E}[R+\gamma v_\pi(S^\prime)|S=s] - [r+\gamma v_\pi(s^\prime)] \Big)}_{\eta} \end{aligned} g~(vπ(s),η)=vπ(s)(r+γvπ(s))=g(vπ(s)) vπ(s)E[R+γvπ(S)S=s]+η (E[R+γvπ(S)S=s][r+γvπ(s)])

Hence, according to the Robbins-Monro algorithm, we can get the TD learning algorithm as

v k + 1 ( s ) = v k ( s ) − α k ( v k ( s ) − [ r k + γ v π ( s k ′ ) ] ) v_{k+1}(s) = v_k(s) - \alpha_k \Big(v_k(s) - [r_k+\gamma v_\pi(s^\prime_k)] \Big) vk+1(s)=vk(s)αk(vk(s)[rk+γvπ(sk)])

We do some modification in order to remove some assumptions of TD learning. One modification is the sample data { ( s , r k , s k ′ ) } \{(s,r_k,s_k^\prime) \} {(s,rk,sk)} is changed to { ( s t , r t + 1 , s t + 1 ) } \{(s_t, r_{t+1}, s_{t+1}) \} {(st,rt+1,st+1)}. Due to the modification the algorithm is called temporal-difference learning. Rewrite it in a more concise way:

TD learning : { v t + 1 ( s t ) ⏟ new estimation = v t ( s t ) ⏟ current estimation − α t ( s t ) [ v t ( s t ) − ( r t + 1 + γ v t ( s t + 1 ) ) ⏟ TD target  v ˉ t ] ⏞ TD error or Innovation  δ t v t + 1 ( s ) = v t ( s ) , for all  s ≠ s t \text{TD learning} : \left \{ \begin{aligned} \textcolor{red}{\underbrace{v_{t+1}(s_t)}_{\text{new estimation}}} & \textcolor{red}{= \underbrace{v_t(s_t)}_{\text{current estimation}} - \alpha_t(s_t) \overbrace{\Big[v_t(s_t) - \underbrace{(r_{t+1} +\gamma v_t(s_{t+1}))}_{\text{TD target } \bar{v}_t} \Big]}^{\text{TD error or Innovation } \delta_t}} \\ \textcolor{red}{v_{t+1}(s)} & \textcolor{red}{= v_t(s)}, \quad \text{for all } s\ne s_t \end{aligned} \right. TD learning: new estimation vt+1(st)vt+1(s)=current estimation vt(st)αt(st)[vt(st)TD target vˉt (rt+1+γvt(st+1))] TD error or Innovation δt=vt(s),for all s=st

where t = 0 , 1 , 2 , … t=0,1,2,\dots t=0,1,2,. Here, v t ( s t ) v_t(s_t) vt(st) is the estimated state value of v π ( s t ) v_\pi(s_t) vπ(st) and a t ( s t ) a_t(s_t) at(st) is the learning rate of s t s_t st at time t t t. And

v ˉ t ≜ r t + 1 + γ v ( s t + 1 ) \bar{v}_t \triangleq r_{t+1}+\gamma v(s_{t+1}) vˉtrt+1+γv(st+1)

is called the TD target and

δ t ≜ v ( s t ) − [ r t + 1 + γ v ( s t + 1 ) ] = v ( s t ) − v ˉ t \delta_t \triangleq v(s_t) - [r_{t+1}+\gamma v(s_{t+1})] = v(s_t) - \bar{v}_t δtv(st)[rt+1+γv(st+1)]=v(st)vˉt

is called the TD error. TD error reflects the deficiency between the current estimate v t v_t vt and the true state value v π v_\pi vπ.

7.2 TD learning of action value: Sarsa

​ Sarsa is an algorithm directly estimate action values. Estimating action value is important because the policy can be improved based on action values.

​ Recall the Bellman equation of action value in section 2.5

q π ( s , a ) = ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v π ( s ′ ) (elementwise form) = ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) ∑ a ′ ∈ A ( s ′ ) π ( a ′ ∣ s ′ ) q π ( s ′ , a ′ ) (elementwise form) \begin{aligned} q_\pi(s,a) & =\sum_r p(r|s,a)r + \gamma\sum_{s^\prime} p(s^\prime|s,a) v_\pi(s^\prime) \quad \text{(elementwise form)} \\ & = \sum_r p(r|s,a)r + \gamma\sum_{s^\prime} p(s^\prime|s,a) \sum_{a^\prime \in \mathcal{A}(s^\prime)}\pi(a^\prime|s^\prime) q_\pi(s^\prime,a^\prime) \quad \text{(elementwise form)} \end{aligned} qπ(s,a)=rp(rs,a)r+γsp(ss,a)vπ(s)(elementwise form)=rp(rs,a)r+γsp(ss,a)aA(s)π(as)qπ(s,a)(elementwise form)

Due to the conditional probability p ( a , b ) = p ( b ) p ( a ∣ b ) p(a,b)=p(b)p(a|b) p(a,b)=p(b)p(ab), we have

p ( s ′ , a ′ ∣ s , a ) = p ( s ′ ∣ s , a ) p ( a ′ ∣ s ′ , s , a ) (conditional probility) = p ( s ′ ∣ s , a ) p ( a ′ ∣ s ′ ) (due to conditional independence) = p ( s ′ ∣ s , a ) π ( a ′ ∣ s ′ ) \begin{aligned} p(s^\prime, a^\prime |s,a) & = p(s^\prime|s,a)p(a^\prime|s^\prime, s, a) \quad \text{(conditional probility)} \\ & = p(s^\prime|s,a) p(a^\prime|s^\prime) \quad \text{(due to conditional independence)} \\ & = p(s^\prime|s,a) \pi(a^\prime|s^\prime) \end{aligned} p(s,as,a)=p(ss,a)p(as,s,a)(conditional probility)=p(ss,a)p(as)(due to conditional independence)=p(ss,a)π(as)

Due to the above equation, we have

q π ( s , a ) = ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ ∑ a ′ p ( s ′ , a ′ ∣ s , a ) q π ( s ′ , a ′ ) q_\pi(s,a) = \sum_r p(r|s,a)r + \gamma \sum_{s^\prime} \sum_{a^\prime} p(s^\prime,a^\prime|s,a) q_\pi(s^\prime,a^\prime) qπ(s,a)=rp(rs,a)r+γsap(s,as,a)qπ(s,a)

Regard the probability p ( r ∣ s , a ) p(r|s,a) p(rs,a) and p ( s ′ , a ′ ∣ s , a ) p(s^\prime, a^\prime |s,a) p(s,as,a) as the distribution of random variable R R R and S ′ S^\prime S respectively. Then rewrite above equation into expectation form

q π ( s , a ) = E [ R + γ q π ( S ′ , A ′ ) ∣ S = s , A = a ] ,  for all  s , a (expectation form) \textcolor{red}{ q_\pi(s,a) = \mathbb{E}\Big[ R + \gamma q_\pi(S^\prime,A^\prime) \Big| S=s, A=a\Big] }, \text{ for all }s,a \quad \text{(expectation form)} qπ(s,a)=E[R+γqπ(S,A) S=s,A=a], for all s,a(expectation form)

where R , S , S ′ R,S,S^\prime R,S,S are random variables, denote immediate reward, currnet state and next state respectively.

Hence, we can use the Robbins-Monro algorithm to sovle the Bellamn eqaution of action value. We can define

g ( q π ( s , a ) ) ≜ q π ( s , a ) − E [ R + γ q π ( S ′ , A ′ ) ∣ S = s , A = a ] g(q_\pi(s,a)) \triangleq q_\pi(s,a) - \mathbb{E}{\Big[ R + \gamma q_\pi(S^\prime,A^\prime) \Big| S=s, A=a\Big]} g(qπ(s,a))qπ(s,a)E[R+γqπ(S,A) S=s,A=a]

We only can get the observation with noise that

g ~ ( q π ( s , a ) , η ) = q π ( s , a ) − [ r + γ q π ( s ′ , a ′ ) ] = q π ( s , a ) − E [ R + γ q π ( S ′ , A ′ ) ∣ S = s , A = a ] ⏟ g ( q π ( s , a ) ) + [ E [ R + γ q π ( S ′ , A ′ ) ∣ S = s , A = a ] − [ r + γ q π ( s ′ , a ′ ) ] ] ⏟ η \begin{aligned} \tilde{g}\Big(q_\pi(s,a),\eta \Big) & = q_\pi(s,a) - \Big[r+\gamma q_\pi(s^\prime,a^\prime) \Big] \\ & = \underbrace{q_\pi(s,a) - \mathbb{E}{\Big[ R + \gamma q_\pi(S^\prime,A^\prime)\Big| S=s, A=a\Big]}}_{g(q_\pi(s,a))} + \underbrace{\Bigg[\mathbb{E}{\Big[ R + \gamma q_\pi(S^\prime,A^\prime) \Big| S=s, A=a\Big]} - \Big[r+\gamma q_\pi(s^\prime,a^\prime)\Big] \Bigg]}_{\eta} \end{aligned} g~(qπ(s,a),η)=qπ(s,a)[r+γqπ(s,a)]=g(qπ(s,a)) qπ(s,a)E[R+γqπ(S,A) S=s,A=a]+η [E[R+γqπ(S,A) S=s,A=a][r+γqπ(s,a)]]

Hence, according to the Robbins-Monro algorithm, we can get Sarsa as

q k + 1 ( s , a ) = q k ( s , a ) − α k [ q k ( s , a ) − ( r k + γ q k ( s k ′ , a k ′ ) ) ] q_{k+1}(s,a) = q_k(s,a) - \alpha_k \Big[ q_k(s,a) - \big(r_k+\gamma q_k(s^\prime_k,a^\prime_k) \big) \Big] qk+1(s,a)=qk(s,a)αk[qk(s,a)(rk+γqk(sk,ak))]

Similar to the TD learning estimates state value in last section, we do some modification in above equation. The sampled data ( s , a , r k , s k ′ , a k ′ ) (s,a,r_k,s^\prime_k,a^\prime_k) (s,a,rk,sk,ak) is changed to ( s t , a t , r t + 1 , s t + 1 , a t + 1 ) (s_t,a_t,r_{t+1},s_{t+1},a_{t+1}) (st,at,rt+1,st+1,at+1). Hence, the Sarsa becomes

Sarsa : { q t + 1 ( s t , a t ) = q t ( s t , a t ) − α t ( s t , a t ) [ q t ( s t , a t ) − ( r t + 1 + γ q t ( s t + 1 , a t + 1 ) ) ] q t + 1 ( s , a ) = q t ( s , a ) , for all  ( s , a ) ≠ ( s t , a t ) \text{Sarsa} : \left \{ \begin{aligned} \textcolor{red}{q_{t+1}(s_t,a_t)} & \textcolor{red}{= q_t(s_t,a_t) - \alpha_t(s_t,a_t) \Big[q_t(s_t,a_t) - (r_{t+1} +\gamma q_t(s_{t+1},a_{t+1})) \Big]} \\ \textcolor{red}{q_{t+1}(s,a)} & \textcolor{red}{= q_t(s,a)}, \quad \text{for all } (s,a) \ne (s_t,a_t) \end{aligned} \right. Sarsa: qt+1(st,at)qt+1(s,a)=qt(st,at)αt(st,at)[qt(st,at)(rt+1+γqt(st+1,at+1))]=qt(s,a),for all (s,a)=(st,at)

where t = 0 , 1 , 2 , … t=0,1,2,\dots t=0,1,2,. Here, q t ( s , a t ) q_t(s,a_t) qt(s,at) is the estimated action value of ( s t , a t ) (s_t,a_t) (st,at); α t ( s t , a t ) \alpha_t(s_t,a_t) αt(st,at) is the learning rate depending on s t , a t s_t,a_t st,at.

​ Sarsa is nothing but an action-value version of the TD algorithm. Sarsa is also implemented with policy improvement steps such as ϵ \epsilon ϵ- greedy algorithm. There is a point that should be noticed. In the q q q-value update step, unlike the model-based policy iteration or value iteration algorithm where the values of all states are updated in each iteration, Sarsa only updates a single state-action pair that is visited at time step t t t.

Pesudocode:

Image

7.3 TD learning of action value: Expected Sarsa

​ Recall the Bellman equation of action value

q π ( s , a ) = ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v π ( s ′ ) (elementwise form) q_\pi(s,a) = \sum_r p(r|s,a)r + \gamma\sum_{s^\prime} p(s^\prime|s,a) v_\pi(s^\prime) \quad \text{(elementwise form)} qπ(s,a)=rp(rs,a)r+γsp(ss,a)vπ(s)(elementwise form)

Regard the probability p ( r ∣ s , a ) p(r|s,a) p(rs,a) and p ( s ′ ∣ s , a ) p(s^\prime|s,a) p(ss,a) as the distribution of random variable R R R and v π ( S ′ ) v_\pi(S^\prime) vπ(S). Then, we have the expectation form of Bellman equation of action value.

q π ( s , a ) = E [ R + γ v π ( S ′ ) ∣ S = s , A = a ] (expectation form) ( 1 ) q_\pi(s,a) = \mathbb{E}[R + \gamma v_\pi(S^\prime)|S=s,A=a] \quad \text{(expectation form)} \quad (1) qπ(s,a)=E[R+γvπ(S)S=s,A=a](expectation form)(1)

According to the definition of state value we have

E [ q π ( s , A ) ∣ s ] = ∑ a ∈ A ( s ) π ( a ∣ s ) q π ( s , a ) = v π ( s ) → E [ q π ( S ′ , A ) ∣ S ′ ] = v π ( S ′ ) ( 2 ) \begin{aligned} \mathbb{E}[q_\pi(s, A) | s] & = \sum_{a\in\mathcal{A}(s)} \pi(a|s) q_\pi(s,a) = v_\pi(s) \\ \to \mathbb{E}[q_\pi(S^\prime, A) | S^\prime] & = v_\pi(S^\prime) \quad (2) \end{aligned} E[qπ(s,A)s]E[qπ(S,A)S]=aA(s)π(as)qπ(s,a)=vπ(s)=vπ(S)(2)

Subtitute ( 2 ) (2) (2) into ( 1 ) (1) (1) we have

q π ( s , a ) = E [ R + γ E [ q π ( S ′ , A ) ∣ S ′ ] ∣ S = s , A = a ] ,  for all  s , a (expectation form) \textcolor{red}{q_\pi(s,a) = \mathbb{E} \Big[ R+\gamma \mathbb{E} \big[ q_\pi(S^\prime, A)|S^\prime \big] \Big| S=s, A=a \Big]}, \text{ for all }s,a \quad \text{(expectation form)} qπ(s,a)=E[R+γE[qπ(S,A)S] S=s,A=a], for all s,a(expectation form)

Rewirte it into root finding parttern

g ( q π ( s , a ) ) ≜ q π ( s , a ) − E [ R + γ E [ q π ( S ′ , A ) ∣ S ′ ] ∣ S = s , A = a ] g(q_\pi(s,a)) \triangleq q_\pi(s,a) - \mathbb{E} \Big[ R+\gamma \mathbb{E} \big[ q_\pi(S^\prime, A)|S^\prime \big] \Big| S=s, A=a \Big] g(qπ(s,a))qπ(s,a)E[R+γE[qπ(S,A)S] S=s,A=a]

We can only get the observation with noise η \eta η

g ~ ( q π ( s , a ) , η ) = q π ( s , a ) − ( r + γ E [ q π ( s ′ , A ) ∣ s ′ ] ) = q π ( s , a ) − E [ R + γ E [ q π ( S ′ , A ) ∣ S ′ ] ∣ S = s , A = a ] ⏟ g ( q π ( s , a ) ) + E [ R + γ E [ q π ( S ′ , A ) ∣ S ′ ] ∣ S = s , A = a ] − ( r + γ E [ q π ( s ′ , A ) ∣ s ′ ] ) ⏟ η \begin{aligned} \tilde{g}(q_\pi(s,a), \eta) & = q_\pi(s,a) - \Big(r + \gamma \mathbb{E} \big[ q_\pi(s^\prime, A)|s^\prime \big] \Big) \\ & = \underbrace{q_\pi(s,a) - \mathbb{E} \Big[ R+\gamma \mathbb{E} \big[ q_\pi(S^\prime, A)|S^\prime \big] \Big| S=s, A=a \Big]}_{g(q_\pi(s,a))} + \underbrace{\mathbb{E} \Big[ R+\gamma \mathbb{E} \big[ q_\pi(S^\prime, A)|S^\prime \big] \Big| S=s, A=a \Big] - \Big(r + \gamma \mathbb{E} \big[ q_\pi(s^\prime, A)|s^\prime \big] \Big)}_{\eta} \end{aligned} g~(qπ(s,a),η)=qπ(s,a)(r+γE[qπ(s,A)s])=g(qπ(s,a)) qπ(s,a)E[R+γE[qπ(S,A)S] S=s,A=a]+η E[R+γE[qπ(S,A)S] S=s,A=a](r+γE[qπ(s,A)s])

Hence, we can implement Robbins-Monro algorithm to find the root of g ( q π ( s , a ) ) g(q_\pi(s,a)) g(qπ(s,a)) that

q k + 1 ( s , a ) = q k ( s , a ) − α k ( s , a ) [ q k ( s , a ) − ( r k + γ E [ q k ( s k ′ , A ) ∣ s k ′ ] ) ] q_{k+1}(s,a) = q_k(s,a) - \alpha_k(s,a) \Bigg[ q_k(s,a) - \Big(r_k + \gamma \mathbb{E} \big[ q_k(s^\prime_k, A)|s^\prime_k \big] \Big) \Bigg] qk+1(s,a)=qk(s,a)αk(s,a)[qk(s,a)(rk+γE[qk(sk,A)sk])]

Similar to the TD learning estimates state value, we do some modification in above equation. The sampled data ( s , a , r k , s k ′ ) (s,a,r_k,s^\prime_k) (s,a,rk,sk) is changed to ( s t , a t , r t + 1 , s t + 1 ) (s_t,a_t,r_{t+1},s_{t+1}) (st,at,rt+1,st+1). Hence, the Expected-Sarsa becomes

Expected-Sarsa : { q t + 1 ( s t , a t ) = q t ( s t , a t ) − α t ( s t , a t ) [ q t ( s t , a t ) − ( r t + 1 + γ E [ q t ( s t + 1 , A ∣ s t + 1 ) ] ) ] q t + 1 ( s , a ) = q t ( s , a ) , for all  ( s , a ) ≠ ( s t , a t ) \text{Expected-Sarsa} : \left \{ \begin{aligned} \textcolor{red}{q_{t+1}(s_t,a_t)} & \textcolor{red}{= q_t(s_t,a_t) - \alpha_t(s_t,a_t) \Big[q_t(s_t,a_t) - \big( r_{t+1} +\gamma \mathbb{E}[q_t(s_{t+1},A|s_{t+1})] \big) \Big]} \\ \textcolor{red}{q_{t+1}(s,a)} & \textcolor{red}{= q_t(s,a)}, \quad \text{for all } (s,a) \ne (s_t,a_t) \end{aligned} \right. Expected-Sarsa: qt+1(st,at)qt+1(s,a)=qt(st,at)αt(st,at)[qt(st,at)(rt+1+γE[qt(st+1,Ast+1)])]=qt(s,a),for all (s,a)=(st,at)

7.4 TD learning of action values: n n n-step Sarsa

​ Recall the definition of action value is

q π ( s , a ) = E [ G t ∣ S t = s , A t = a ] q_\pi(s,a) = \mathbb{E}[G_t|S_t=s, A_t=a] qπ(s,a)=E[GtSt=s,At=a]

The discounted return G t G_t Gt can be written in different forms as

Sarsa ⟵ G t ( 1 ) = R t + 1 + γ q π ( S t + 1 , A t + 1 ) G t ( 2 ) = R t + 1 + γ R t + 2 + γ 2 q π ( S t + 2 , A t + 2 ) ⋮ n -step Sarsa ⟵ G t ( n ) = R t + 1 + γ R t + 2 + ⋯ + γ n q π ( S t + n , A t + n ) ⋮ Monte Carlo ⟵ G t ( ∞ ) = R t + 1 + γ R t + 2 + γ 2 R t + 3 + ⋯ \begin{aligned} \text{Sarsa} \longleftarrow G_t^{(1)} & = R_{t+1} + \gamma q_\pi(S_{t+1},A_{t+1}) \\ G_t^{(2)} & = R_{t+1} + \gamma R_{t+2} + \gamma^2 q_\pi(S_{t+2},A_{t+2}) \\ & \vdots \\ n\text{-step Sarsa} \longleftarrow G_t^{(n)} & = R_{t+1} + \gamma R_{t+2} + \cdots +\gamma^n q_\pi(S_{t+n},A_{t+n}) \\ & \vdots \\ \text{Monte Carlo} \longleftarrow G_t^{(\infty)} & = R_{t+1} + \gamma R_{t+2} + \gamma^2 R_{t+3}+ \cdots \\ \end{aligned} SarsaGt(1)Gt(2)n-step SarsaGt(n)Monte CarloGt()=Rt+1+γqπ(St+1,At+1)=Rt+1+γRt+2+γ2qπ(St+2,At+2)=Rt+1+γRt+2++γnqπ(St+n,At+n)=Rt+1+γRt+2+γ2Rt+3+

It should be noted that G t = G t ( 1 ) = G t ( 2 ) = G t ( n ) = G t ( ∞ ) G_t=G_t^{(1)}=G_t^{(2)}=G_t^{(n)}=G_t^{(\infty)} Gt=Gt(1)=Gt(2)=Gt(n)=Gt(), where the superscripts merely indicate the different decomposition structures of G t G_t Gt.

​ Sarsa aims to solve

q π ( s , a ) = E [ G t ( 1 ) ∣ s , a ] = E [ R t + 1 + γ q π ( S t + 1 , A t + 1 ) ∣ s , a ] q_\pi(s,a) = \mathbb{E}[G_t^{(1)}|s,a] = \mathbb{E} [R_{t+1}+\gamma q_\pi(S_{t+1},A_{t+1})|s,a ] qπ(s,a)=E[Gt(1)s,a]=E[Rt+1+γqπ(St+1,At+1)s,a]

MC learning aims to solve

q π ( s , a ) = E [ G t ( ∞ ) ∣ s , a ] = E [ R t + 1 + γ R t + 2 + γ 2 R t + 3 + ⋯ ∣ s , a ] q_\pi(s,a) = \mathbb{E} [G_t^{(\infty)}|s,a] = \mathbb{E} [R_{t+1}+\gamma R_{t+2} + \gamma^2 R_{t+3}+\cdots |s,a] qπ(s,a)=E[Gt()s,a]=E[Rt+1+γRt+2+γ2Rt+3+s,a]

n n n-step Sarsa aims to solve

q π ( s , a ) = E [ G t ( n ) ∣ s , a ] = E [ R t + 1 + γ R t + 2 + ⋯ + γ n q π ( S t + n , A t + n ) ∣ s , a ] q_\pi(s,a) = \mathbb{E} [G_t^{(n)}|s,a] = \mathbb{E} [R_{t+1}+\gamma R_{t+2} +\cdots + \gamma^n q_\pi(S_{t+n},A_{t+n}) |s,a] qπ(s,a)=E[Gt(n)s,a]=E[Rt+1+γRt+2++γnqπ(St+n,At+n)s,a]

​ The n n n-step Sarsa algorithm is

n -step Sarsa : { q t + 1 ( s t , a t ) = q t ( s t , a t ) − α t ( s t , a t ) [ q t ( s t , a t ) − ( r t + 1 + γ r t + 2 + ⋯ + γ n q t ( s t + n , a t + n ) ) ] q t + 1 ( s , a ) = q t ( s , a ) , for all  ( s , a ) ≠ ( s t , a t ) n\text{-step Sarsa} : \left \{ \begin{aligned} \textcolor{red}{q_{t+1}(s_t,a_t)} & \textcolor{red}{= q_t(s_t,a_t) - \alpha_t(s_t,a_t) \Big[q_t(s_t,a_t) - (r_{t+1}+ \gamma r_{t+2} + \cdots + \gamma^n q_t(s_{t+n},a_{t+n})) \Big]} \\ \textcolor{red}{q_{t+1}(s,a)} & \textcolor{red}{= q_t(s,a)}, \quad \text{for all } (s,a) \ne (s_t,a_t) \end{aligned} \right. n-step Sarsa: qt+1(st,at)qt+1(s,a)=qt(st,at)αt(st,at)[qt(st,at)(rt+1+γrt+2++γnqt(st+n,at+n))]=qt(s,a),for all (s,a)=(st,at)

7.5 TD learning of optimal action values: Q-learning

​ It should be noted that Sarsa can only estimate the action values of a given policy. It must be combined with a policy improvement step to find optimal policies and hence their optimal action values. By contrast, Q-learning can directly estimate optimal action values.

​ Recall the Bellman optimal equation of state value in section 3.2

v ( s ) = max ⁡ π ∑ a ∈ A ( s ) π ( a ∣ s ) [ ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v ( s ′ ) ] v ( s ) = max ⁡ a ∈ A ( s ) [ ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v ( s ′ ) ] \begin{aligned} v(s) & = \max_\pi \sum_{a\in\mathcal{A}(s)} \pi(a|s) \Big[ \sum_r p(r|s,a)r + \gamma \sum_{s^\prime} p(s^{\prime}|s,a) v(s^{\prime}) \Big] \\ v(s) & = \max_{a\in\mathcal{A}(s)} \Big[\sum_r p(r|s,a)r + \gamma \sum_{s^\prime} p(s^\prime|s,a) v(s^\prime) \Big] \end{aligned} v(s)v(s)=πmaxaA(s)π(as)[rp(rs,a)r+γsp(ss,a)v(s)]=aA(s)max[rp(rs,a)r+γsp(ss,a)v(s)]

where v ( s ) ≜ max ⁡ a ∈ A ( s ) q ( s , a ) v(s)\triangleq \max_{a\in\mathcal{A}(s)} q(s,a) v(s)maxaA(s)q(s,a). Hence we have

max ⁡ a ∈ A ( s ) q ( s , a ) = max ⁡ a ∈ A ( s ) [ ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) v ( s ′ ) ] max ⁡ a ∈ A ( s ) q ( s , a ) = max ⁡ a ∈ A ( s ) [ ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) max ⁡ a ∈ A ( s ) q ( s ′ , a ) ] → q ( s , a ) = ∑ r p ( r ∣ s , a ) r + γ ∑ s ′ p ( s ′ ∣ s , a ) max ⁡ a ∈ A ( s ) q ( s ′ , a ) (elementwise form) \begin{aligned} \max_{a\in\mathcal{A}(s)} q(s,a) & = \max_{a\in\mathcal{A}(s)} \Big[\sum_r p(r|s,a)r + \gamma \sum_{s^\prime} p(s^\prime|s,a) v(s^\prime) \Big] \\ \max_{a\in\mathcal{A}(s)} q(s,a) & = \max_{a\in\mathcal{A}(s)} \Big[\sum_r p(r|s,a)r + \gamma \sum_{s^\prime} p(s^\prime|s,a) \max_{a\in\mathcal{A}(s)} q(s^\prime,a) \Big] \\ \to q(s,a) & = \sum_r p(r|s,a)r + \gamma \sum_{s^\prime} p(s^\prime|s,a) \max_{a\in\mathcal{A}(s)} q(s^\prime,a) \quad \text{(elementwise form)} \end{aligned} aA(s)maxq(s,a)aA(s)maxq(s,a)q(s,a)=aA(s)max[rp(rs,a)r+γsp(ss,a)v(s)]=aA(s)max[rp(rs,a)r+γsp(ss,a)aA(s)maxq(s,a)]=rp(rs,a)r+γsp(ss,a)aA(s)maxq(s,a)(elementwise form)

Rewrite it into expectation form

q ( s , a ) = E [ R + γ max ⁡ a ∈ A ( s ) q ( S ′ , a ) ∣ S = s , A = a ] ,  for all  s , a (expectation form) \textcolor{red}{ q(s,a) = \mathbb{E}[R+\gamma \max_{a\in\mathcal{A}(s)} q(S^\prime,a) |S=s,A=a ] }, \text{ for all }s,a \quad \text{(expectation form)} q(s,a)=E[R+γaA(s)maxq(S,a)S=s,A=a], for all s,a(expectation form)

This equation is the Bellman optimal equation expressed in terms of action values.

Rewrite it into

g ( q ( s , a ) ) ≜ q ( s , a ) − E [ R + γ max ⁡ a ∈ A ( S ′ ) q ( S ′ , a ) ∣ S = s , A = a ] g(q(s,a)) \triangleq q(s,a) - \mathbb{E} [R+\gamma \max_{a\in\mathcal{A}(S^\prime)} q(S^\prime,a) |S=s,A=a ] g(q(s,a))q(s,a)E[R+γaA(S)maxq(S,a)S=s,A=a]

We can get the observation with noise

g ~ ( q ( s , a ) ) = q ( s , a ) − [ r + γ max ⁡ a ∈ A ( s ′ ) q ( s ′ , a ) ] = q ( s , a ) − E [ R + γ max ⁡ a ∈ A ( S ′ ) q ( S ′ , a ) ∣ S = s , A = a ] ⏟ g ( q ( s , a ) ) + E [ R + γ max ⁡ a ∈ A ( S ′ ) q ( S ′ , a ) ∣ S = s , A = a ] − [ r + γ max ⁡ a ∈ A ( s ′ ) q ( s ′ , a ) ] ⏟ η \begin{aligned} \tilde{g}(q(s,a)) & = q(s,a) - \Big[r + \gamma \max_{a\in\mathcal{A}(s^\prime)} q(s^\prime,a) \Big] \\ & = \underbrace{q(s,a) - \mathbb{E} [R+\gamma \max_{a\in\mathcal{A}(S^\prime)} q(S^\prime,a) |S=s,A=a ]}_{g(q(s,a))} + \underbrace{\mathbb{E} [R+\gamma \max_{a\in\mathcal{A}(S^\prime)} q(S^\prime,a) |S=s,A=a ] - \Big[r + \gamma \max_{a\in\mathcal{A}(s^\prime)} q(s^\prime,a) \Big]}_{\eta} \end{aligned} g~(q(s,a))=q(s,a)[r+γaA(s)maxq(s,a)]=g(q(s,a)) q(s,a)E[R+γaA(S)maxq(S,a)S=s,A=a]+η E[R+γaA(S)maxq(S,a)S=s,A=a][r+γaA(s)maxq(s,a)]

Hence, we can implement Robbins-Monro algorithm to find the root

q k + 1 ( s , a ) = q k ( s , a ) − α k ( s , a ) [ q k ( s , a ) − ( r k + γ max ⁡ a ∈ A ( s ′ ) q k ( s ′ , a ) ) ] q_{k+1}(s,a) = q_k(s,a) - \alpha_k(s,a) \Big[q_k(s,a) - \big(r_k + \gamma \max_{a\in\mathcal{A}(s^\prime)} q_k(s^\prime,a) \big) \Big] qk+1(s,a)=qk(s,a)αk(s,a)[qk(s,a)(rk+γaA(s)maxqk(s,a))]

Similar to the TD learning estimates state value, we do some modification in above equation. The sampled data ( s , a , r k , s k ′ ) (s,a,r_k,s^\prime_k) (s,a,rk,sk) is changed to ( s t , a t , r t + 1 , s t + 1 ) (s_t,a_t,r_{t+1},s_{t+1}) (st,at,rt+1,st+1). Hence, the Q-learning becomes

Q-learning : { q t + 1 ( s t , a t ) = q t ( s t , a t ) − α t ( s t , a t ) [ q t ( s t , a t ) − ( r t + 1 + γ max ⁡ a ∈ A ( s t + 1 ) q t ( s t + 1 , a ) ) ] q t + 1 ( s , a ) = q t ( s , a ) , for all  ( s , a ) ≠ ( s t , a t ) \text{Q-learning} : \left \{ \begin{aligned} \textcolor{red}{q_{t+1}(s_t,a_t)} & \textcolor{red}{= q_t(s_t,a_t) - \alpha_t(s_t,a_t) \Big[q_t(s_t,a_t) - (r_{t+1}+ \gamma \max_{a\in\mathcal{A}(s_{t+1})} q_t(s_{t+1},a)) \Big]} \\ \textcolor{red}{q_{t+1}(s,a)} & \textcolor{red}{= q_t(s,a)}, \quad \text{for all } (s,a) \ne (s_t,a_t) \end{aligned} \right. Q-learning: qt+1(st,at)qt+1(s,a)=qt(st,at)αt(st,at)[qt(st,at)(rt+1+γaA(st+1)maxqt(st+1,a))]=qt(s,a),for all (s,a)=(st,at)

Off-policy vs on-policy:

​ There exist two policies in a TD learning task: behavior policy and target policy. The behavoir policy is used to generate experience samples. The target policy is constantly updated toward an optimal policy. When the behavior policy is the same as the target policy, such a kind of learning is called on-policy. Otherwise, when they are different, the learning is called off-policy.

​ The advantage of off-policy learning compared to on-policy learning is that it can search for optimal policies based on the experiences generated by any other policies.

How to determine a algorithm is on-policy or off-policy? The basic reason is that if the algorithm is sovling Bellman equation, then it’s on-policy. This is because Bellman equation is finding the state value or action value under a given policy. Else if the algorithm is sovling Bellman optimal equation, then it’s off-policy. This because Bellman equation does not include any policy, hence, the behavior policy and target policy can be different.

Online learning vs offline learning:

Online learning refers to the case where the value and policy can be updated once an experience sample is obtained. Offline learning refers to the case that the update can only be done after all experience samples have been collected. For example, TD learning is online whereas Monte Carlo learning is offline.

Persudocode:

(On-policy version)

Image

(Off-policy version)

Image

Reference

赵世钰老师的课程

本文来自互联网用户投稿,该文观点仅代表作者本人,不代表本站立场。本站仅提供信息存储空间服务,不拥有所有权,不承担相关法律责任。如若转载,请注明出处:http://www.coloradmin.cn/o/797071.html

如若内容造成侵权/违法违规/事实不符,请联系多彩编程网进行投诉反馈,一经查实,立即删除!

相关文章

Linux内核与内核空间是什么关系呢?

对内核空间的认识清晰了许多。要理解用户空间与内核空间需要有如下的几个认识: 内核的认识:从2个不同的角度来理解,一个是静态的角度,如“芦中人”所比喻,内核可以看做是一个lib库,内核对外提供的API打包…

快速远程桌面控制公司电脑远程办公

文章目录 快速远程桌面控制公司电脑远程办公**第一步****第二步****第三步** 快速远程桌面控制公司电脑远程办公 远程办公的概念很早就被提出来,但似乎并没有多少项目普及落实到实际应用层面,至少在前几年,远程办公距离我们仍然很遥远。但20…

1分钟上手Apifox

1、客户端右上角账号设置-生成令牌 2、IDEA下载插件 Apifox Helper 3、 配置ApiFoxHelper 令牌 4、在controller类界面右键 5、输入项目id 6、项目ID从客户端 项目设置-项目ID获取 7、导入成功 8、右键刷新查看导入的接口 9、自动生成数据(某postman还要自己手输&a…

说一说java中的自定义注解之设计及实现

一、需求背景 比如我们需要对系统的部分接口进行token验证,防止对外的接口裸奔。所以,在调用这类接口前,先校验token的合法性,进而得到登录用户的userId/role/authority/tenantId等信息;再进一步对比当前用户是否有权…

MyBatis 快速入门【中】

😀前言 本篇博文是MyBatis(简化数据库操作的持久层框架)–快速入门[上]的核心部分,分享了MyBatis实现sql的xml配置和一些关联配置、异常分析 🏠个人主页:晨犀主页 🧑个人简介:大家好,我是晨犀&a…

软件外包开发的需求分析

需求分析是软件开发中的关键步骤,其目的是确定用户需要什么样的软件,以及软件应该完成哪些任务。需求分析是软件工程的早期工作,也是软件项目成功的基础,因此花费大量精力和时间去做好需求分析是值得的。今天和大家分享软件需求分…

数字孪生-数字城市效果实现方法

数字孪生-数字城市效果实现方法 效果图: 一、效果分析: .0 1、城市非主展示区域白模快速生成方案: 参考视频: 1、CityEngine 引用数据源生成。 cityengine2022一键生成城市模型,不用再用blendergis_哔哩哔哩_bil…

Python运算符列表及其优先顺序、结合性

本文表格对Python中运算符的优先顺序进行了总结,从最高优先级(最先绑定)到最低优先级(最后绑定)。相同单元格内的运算符具有相同优先级。除非句法显式地给出,否则运算符均指二元运算。 相同单元格内的运算…

数据安全之全景图系列——数据分类分级落地实践

1、数据分类分级现状 我们正处于一个数据爆炸式增长的时代,随着产业数字化转型升级的推进,数据已被国家层面纳入生产要素,并且成为企业、社会和国家层面重要的战略资源。数据分类分级管理不仅是加强数据交换共享、提升数据资源价值的前提条件…

Unreal MorphTarget Connect Bone MetaData Curve功能学习

MorphTarget Connected Bone和MetaData Curve是两个较冷门功能,近期在制作一些功能时留意到这2个内容,故研究一下。 1.MorphTarget Connected Bone 在骨架编辑面板中,选中MorphTarget时,可找到Connected Bone选项: …

[数据库]对数据库事务进行总结

文章目录 1、什么是事务2、事务的特性(ACID)3、并发事务带来的问题4、四个隔离级别: 1、什么是事务 事务是逻辑上的一组操作,要么都执行,要么都不执行。 事务最经典也经常被拿出来说例子就是转账了。假如小明要给小红…

图解SQL基础知识,小白也能看懂的SQL文章

本文介绍关系数据库的设计思想: 在 SQL 中,一切皆关系。 在计算机龄域有许多伟大的设计理念和思想,例如: 在 Unix 中,一切皆文件。 在面向对象的编程语言中,一切皆对象。 关系数据库同样也有自己的设计…

mybatis_使用注解开发

第一步&#xff1a;使用注解写一个接口 Select("select * from user")List<User> getUsers(); 第二步&#xff1a;绑定接口 第三步&#xff1a;测试 官方提示&#xff1a; 使用注解来映射简单语句会使代码显得更加简洁&#xff0c;但对于稍微复杂一点的语句&…

Spring AOP (面向切面编程)原理与代理模式—实例演示

一、AOP介绍和应用场景 Spring 中文文档 (springdoc.cn) Spring | Home 官网 1、AOP介绍&#xff08;为什么会出现AOP &#xff1f;&#xff09; Java是一个面向对象&#xff08;OOP&#xff09;的语言&#xff0c;但它有一些弊端。虽然使用OOP可以通过组合或继承的方…

使用网络 IP 扫描程序的原因

随着网络不断扩展以满足业务需求&#xff0c;高级 IP 扫描已成为网络管理员确保网络可用性和性能的关键任务。在大型网络中扫描 IP 地址可能具有挑战性&#xff0c;这些网络通常包括具有动态 IP、多个 DNS、DHCP 配置和复杂子网的有线和无线设备。使用可提供全面 IP 地址管理 &…

哈尔滨的全年平均气温和降水特点

哈尔滨全年平均气温和降水特点 哈尔滨是我国最靠北的省会城市&#xff0c;冬季那边特别冷&#xff0c;但夏季比较凉爽&#xff0c;下面根据中国气象台的1981-2010年的统计数据[1]&#xff0c;分析哈尔滨全年平均气温和降水的特点有以下几点&#xff1a; 1.全年最高气温在7月&…

ubuntu18.04安装autoware1.15

目录 前言一、准备工作1.安装autoware1.152.安装依赖3.把src/autoware/common/autoware_build_flags/cmake文件夹下的CUDA版本改为11.4&#xff08;或者你电脑上的版本&#xff09; 二、解决报错错误类型1错误类型2错误类型3错误类型4错误类型5错误类型6 前言 本文参考链接&am…

如何降低TCP在局域网环境下的数据传输延迟

以Ping为例。本案例是一个测试题目&#xff0c;只有现象展示&#xff0c;不含解决方案。 ROS_Kinetic_26 使用rosserial_windows实现windows与ROS master发送与接收消息_windows 接收ros1 消息 什么是ping&#xff1f; AI&#xff1a; ping是互联网控制消息协议&#xff08;…

数学建模-预测模型 神经网络

设置测试集&#xff0c;算sse&#xff0c;进行过拟合检验

无涯教程-jQuery - jQuery.get( url, data, callback, type )方法函数

jQuery.get(url&#xff0c;[data]&#xff0c;[callback]&#xff0c;[type])方法使用GET HTTP请求从服务器加载数据。 该方法返回XMLHttpRequest对象。 jQuery.get( url, [data], [callback], [type] ) - 语法 $.get( url, [data], [callback], [type] ) 这是此方法使用的…