Control and Observation for Stochastic Partial Differential Equations
Transcription
Control and Observation for Stochastic Partial Differential Equations
Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Control and Observation for Stochastic Partial Differential Equations Qi Lü LJLL, UPMC October 5, 2013 Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Do the controllability problems for stochastic ordinary differential equations(SDEs for short) and stochastic partial differential equations deserve to be studied? Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Do the controllability problems for stochastic ordinary differential equations(SDEs for short) and stochastic partial differential equations deserve to be studied? • Yes. I can find many evidences to support my conclusion. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Do the controllability problems for stochastic ordinary differential equations(SDEs for short) and stochastic partial differential equations deserve to be studied? • Yes. I can find many evidences to support my conclusion. • I only point out that, in the control theory of deterministic ODEs and PDEs, both the optimal control problems and the controllability problems are studied extensively. However, for the stochastic counterpart, the situation is quite different. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Many excellent mathematicians have left their footprints in the stochastic optimal control theory. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Many excellent mathematicians have left their footprints in the stochastic optimal control theory. • A very brief list includes J.-L. Lions, A. Bensoussan, W. H. Fleming, J.-M. Bismut, P.-L. Lions, etc. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Advertisement • Many excellent mathematicians have left their footprints in the stochastic optimal control theory. • A very brief list includes J.-L. Lions, A. Bensoussan, W. H. Fleming, J.-M. Bismut, P.-L. Lions, etc. • The stochastic controllability problems attract very few mathematicians. It seems that this area is almost an uncultivated land. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controllability results for SPDEs and open problems Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 1. A toy model • Let us recall some basic result for deterministic ODE. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 1. A toy model • Let us recall some basic result for deterministic ODE. • Consider the following controlled system: dy = Ay + Bu, dt y (0) = y0 . t ∈ (0, T ), (1) where A ∈ Rn×n , B ∈ Rn×m , T > 0. System (1) is said to be exactly controllable on (0, T ) if for any y0 , y1 ∈ Rn , there exists a u ∈ L2 (0, T ; Rm ) such that y (T ) = y1 . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 1. A toy model • Let us recall some basic result for deterministic ODE. • Consider the following controlled system: dy = Ay + Bu, dt y (0) = y0 . t ∈ (0, T ), (1) where A ∈ Rn×n , B ∈ Rn×m , T > 0. System (1) is said to be exactly controllable on (0, T ) if for any y0 , y1 ∈ Rn , there exists a u ∈ L2 (0, T ; Rm ) such that y (T ) = y1 . µSystem (1) is exactly controllable on (0, T )⇔ • Theorem 1 rank(B, AB, A2 B, · · · , An−1 B) = n. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Consider a linear stochastic differential equation: ( dy = Ay + Bu dt + CydW (t), y (0) = y0 ∈ Rn , t ≥ 0, (2) where A, C ∈ Rn×n and B ∈ Rn×m , {W (t)}t≥0 is a standard one dimensional Brownian motion. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Consider a linear stochastic differential equation: ( dy = Ay + Bu dt + CydW (t), t ≥ 0, y (0) = y0 ∈ Rn , (2) where A, C ∈ Rn×n and B ∈ Rn×m , {W (t)}t≥0 is a standard one dimensional Brownian motion. • System (2) is said to be exactly controllable if for any y0 ∈ Rn and yT ∈ L2FT (Ω; Rn ), there exists a control u(·) ∈ L2F (Ω; L2 (0, T ; Rm )) such that the corresponding solution y (·) of (2) satisfies y (T ) = yT . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • u(·) ∈ L2F (Ω; L2 (0, T ; Rm )) means that one can only utilize the information of the noise {W (s)}0≤s≤t which can be observed at time t. This is a key point in the study of stochastic control problems. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • u(·) ∈ L2F (Ω; L2 (0, T ; Rm )) means that one can only utilize the information of the noise {W (s)}0≤s≤t which can be observed at time t. This is a key point in the study of stochastic control problems. • L2F (Ω; Rn ) is the natural state space at time T when there is T noise. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • u(·) ∈ L2F (Ω; L2 (0, T ; Rm )) means that one can only utilize the information of the noise {W (s)}0≤s≤t which can be observed at time t. This is a key point in the study of stochastic control problems. • L2F (Ω; Rn ) is the natural state space at time T when there is T noise. • In virtue of a result by S. Peng (1994), system (2) is NOT exactly controllable for any B and any m! Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Thus, we consider the following alternative system: ( dy = Ay + Bu dt + (Cy + Dv )dW (t), t ≥ 0, y (0) = y0 ∈ Rn , where A, C , D ∈ Rn×n , B ∈ Rn×m , u ∈ L2F (0, T ; Rm ) and v ∈ L2F (0, T ; Rn ). (3) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Thus, we consider the following alternative system: ( dy = Ay + Bu dt + (Cy + Dv )dW (t), t ≥ 0, y (0) = y0 ∈ Rn , where A, C , D ∈ Rn×n , B ∈ Rn×m , u ∈ L2F (0, T ; Rm ) and v ∈ L2F (0, T ; Rn ). µSystem (4) is exactly controllable on (0, T )⇔ • Theorem 2 rank(B, AB, A2B, · · · , An−1 B) = n and rank(D) = n. (3) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Is Theorem 2 a trivial corollary of Theorem 1? Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Is Theorem 2 a trivial corollary of Theorem 1? • Taking v (t) = D −1 Cy (t), we get ( dy = Ay + Bu dt, t ≥ 0, y (0) = y0 ∈ Rn . Is this system exactly controllable when (A, B) satisfies the Kalman rank condition? NO. (4) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Is Theorem 2 a trivial corollary of Theorem 1? • Taking v (t) = D −1 Cy (t), we get ( dy = Ay + Bu dt, t ≥ 0, y (0) = y0 ∈ Rn . Is this system exactly controllable when (A, B) satisfies the Kalman rank condition? NO. • The reason is that we expect the final state could be any element in L2FT (Ω; Rn ). (4) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • How to prove Theorem 2? Following the idea for deterministic system, we should study the observability estimate for the adjoint system of (2). Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • How to prove Theorem 2? Following the idea for deterministic system, we should study the observability estimate for the adjoint system of (2). • The adjoint system of (2) reads: ( dz = − A⊤ z + C ⊤ Z dt + ZdW , z(T ) = zT ∈ L2FT (Ω; Rn ). t ∈ [0, T ], (5) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • How to prove Theorem 2? Following the idea for deterministic system, we should study the observability estimate for the adjoint system of (2). • The adjoint system of (2) reads: ( dz = − A⊤ z + C ⊤ Z dt + ZdW , t ∈ [0, T ], z(T ) = zT ∈ L2FT (Ω; Rn ). (5) • Equation (5) is a backward stochastic ODE. Such kind of equations were first introduced by J. Bismut in 1970s and studied extensively in the last twenty years by E. Pardoux, S. Peng, N. El Karoui, R. Buckdahn, X. Zhou, J. Yong, etc. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We only need to prove the following inequality: |zT |2L2 FT (Ω;Rn ) ≤ CE Z 0 T (|B ⊤ z(t)|2 + Z 2 )dt. (6) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We only need to prove the following inequality: |zT |2L2 FT (Ω;Rn ) ≤ CE Z T (|B ⊤ z(t)|2 + Z 2 )dt. (6) 0 • Please note that we now deal with infinite dimensional space. One cannot simply mimic the proof for observability estimate for ODE to obtain (6). For example, in this case, unique continuation does not imply the observability estimate. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 2. Exact controllability of Stochastic transport equations • Let T > 0 and G ⊂ Rd (d ∈ N) a strictly convex bounded domain with the C 1 boundary Γ. Let U ∈ Rd with |U|Rd = 1. Denote by Γ−S = {x ∈ Γ : U · ν(x) ≤ 0}, Γ+S = Γ \ Γ−S . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 2. Exact controllability of Stochastic transport equations • Let T > 0 and G ⊂ Rd (d ∈ N) a strictly convex bounded domain with the C 1 boundary Γ. Let U ∈ Rd with |U|Rd = 1. Denote by Γ−S = {x ∈ Γ : U · ν(x) ≤ 0}, Γ+S = Γ \ Γ−S . • Consider a linear stochastic differential equation: dy +U ·∇ydt = a1 ydt + a2 y +v dW (t) y =u y (0) = y0 Here y0 ∈ L2 (G ), a1 , a2 ∈ in (0, T )×G , on (0, T )×Γ−S , in G . ∞ L∞ F (0, T ; L (G )). (7) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • A solution to the system (7) is a stochastic process y ∈ L2F (Ω; C ([0, T ]; L2 (G ))) such that for every τ ∈ [0, T ] and every φ ∈ C 1 (G ), φ = 0 on Γ+S , it holds that Z Z y (τ, x)φ(x)dx − y0 (x)φ(x)dx G G Z Z Z τZ τ − y (s,x)U ·∇φ(x,U)dxds+ u(s,x)φ(x)U ·νdΓ−S ds = 0 G τ Z Z 0 + Z 0 0 Γ−S a1 (s, x)y (s, x)φ(x)dxds G τZ G a2 (s,x)y (s,x)+v (s,x) φ(x)dxdW (s), P-a.s. (8) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • For each y0 ∈ L2 (G ), the system (7) admits a unique solution y . Further, there is a constant C > 0 such that for every y0 , it holds that |y |L2 2 F (Ω;C ([0,T ];L (G ))) ≤ e Cr1 E|y0 |L2 (G ) + |u|L2 2 F (0,T ;Lw (Γ−S )) + |v |L2 2 F (0,T ;L (G )) Here r1 = |a1 |2L∞ (0,T ;L∞ (G )) + |a2 |2L∞ (0,T ;L∞ (G )) + 1. F F . (9) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • System (7) is said to be exactly controllable at time T if for every initial state y0 ∈ L2 (G ) and every y1 ∈ L2FT (Ω; L2 (G )), one can find a pair of controls (u, v ) ∈ L2F (0, T ; L2w (Γ−S )) × L2F (0, T ; L2 (G )) such that the solution y of the system (7) satisfies that y (T ) = y1 , P-a.s. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • System (7) is said to be exactly controllable at time T if for every initial state y0 ∈ L2 (G ) and every y1 ∈ L2FT (Ω; L2 (G )), one can find a pair of controls (u, v ) ∈ L2F (0, T ; L2w (Γ−S )) × L2F (0, T ; L2 (G )) such that the solution y of the system (7) satisfies that y (T ) = y1 , P-a.s. • Put R = max |x1 − x2 |Rd . x1 ,x2 ∈G Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • System (7) is said to be exactly controllable at time T if for every initial state y0 ∈ L2 (G ) and every y1 ∈ L2FT (Ω; L2 (G )), one can find a pair of controls (u, v ) ∈ L2F (0, T ; L2w (Γ−S )) × L2F (0, T ; L2 (G )) such that the solution y of the system (7) satisfies that y (T ) = y1 , P-a.s. • Put R = max |x1 − x2 |Rd . x1 ,x2 ∈G • Theorem 3: System (7) is exactly controllable at time T , provided that T > R. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We put two controls on the system. Moreover, the control v acts on the whole domain. Compared with the deterministic transport solution, it seems that our choice of controls is too restrictive. One may consider the following four weaker cases for designing the control. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We put two controls on the system. Moreover, the control v acts on the whole domain. Compared with the deterministic transport solution, it seems that our choice of controls is too restrictive. One may consider the following four weaker cases for designing the control. • 1. Only one control is acted on the system, that is, u = 0 or v = 0 in (7). Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We put two controls on the system. Moreover, the control v acts on the whole domain. Compared with the deterministic transport solution, it seems that our choice of controls is too restrictive. One may consider the following four weaker cases for designing the control. • 1. Only one control is acted on the system, that is, u = 0 or v = 0 in (7). • 2. Neither u nor v is zero. But v = 0 in (0, T ) × G0 , where G0 is a nonempty open subset of G . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We put two controls on the system. Moreover, the control v acts on the whole domain. Compared with the deterministic transport solution, it seems that our choice of controls is too restrictive. One may consider the following four weaker cases for designing the control. • 1. Only one control is acted on the system, that is, u = 0 or v = 0 in (7). • 2. Neither u nor v is zero. But v = 0 in (0, T ) × G0 , where G0 is a nonempty open subset of G . • 3. Two controls are acted on the system. But both of them are in the drift term. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Theorem 4: If u ≡ 0 or v ≡ 0 in the system (7), then the system (7) is not exactly controllable at any time T . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Theorem 4: If u ≡ 0 or v ≡ 0 in the system (7), then the system (7) is not exactly controllable at any time T . • Theorem 5: Let G0 be a nonempty open subset of G . If v ≡ 0 (0, T ) × G0 , then the system (7) is not exactly controllable at any time T . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • For the third case, we consider the following controlled equation: dy +U ·∇ydt= a1 y +ℓ dt +a2 ydW (t) y =u y (0) = y0 Here ℓ ∈ L2F (0, T ; L2 (G )) is a control. in (0, T ) × G , on (0, T ) × Γ−S , in G . (10) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • For the third case, we consider the following controlled equation: dy +U ·∇ydt= a1 y +ℓ dt +a2 ydW (t) y =u y (0) = y0 Here ℓ ∈ L2F (0, T ; L2 (G )) in (0, T ) × G , on (0, T ) × Γ−S , in G . (10) is a control. • Theorem 6 The system (10) is not exactly controllable for any T > 0. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Let us introduce the adjoint system: dz +U ·∇zdt= b1 z +b2 Z dt+(b3 z +Z )dW (t) in (0, T )×G , z =0 on (0,T )×Γ+S , z(T ) = zT in G . (11) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Let us introduce the adjoint system: dz +U ·∇zdt= b1 z +b2 Z dt+(b3 z +Z )dW (t) in (0, T )×G , z =0 on (0,T )×Γ+S , z(T ) = zT in G . (11) ∞ • Here zT ∈ L2F (Ω; L2 (G )),b1 , b2 , b3 ∈ L∞ F (0, T ; L (G )). T Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • A solution to (11) is a pair of stochastic processes (z, Z ) ∈ CF ([0, T ]; L2 (Ω; L2 (G ))) × L2F (0, T ; L2 (G )) such that for every ψ ∈ C 1 (G ) with ψ = 0 on Γ−S and arbitrary τ ∈ [0, T ], it holds that Z Z zT (x)ψ(x)dx − z(τ, x)ψ(x)dx G − G Z τ = T Z Z TZ Z TZ τ G τ + z(s, x)U · ∇ψ(x)dxds G G b1 (s,x)z(s,x)+b2 (s,x)Z (s,x) ψ(x)dxds b3 (s,x)z(s,x)+Z (s,x) ψ(x)dxdW (s), P-a.e. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • For any zT ∈ L2 (Ω, FT , P; L2 (G )), the equation (11) admits a unique solution (z, Z ) satisfiing that |z|CF ([0,T ];L2 (Ω;L2 (G ))) +|Z |L2 (0,T ;L2 (G )) ≤ e Cr2 |zT |L2 F where △ r2 = 3 X i =1 FT |bi |4L∞ (0,T ;L∞ (G )) + 1. F (Ω;L2 (G )) , (12) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The observability estimate for (11) reads |zT |L2 FT (Ω;L2 (G )) ≤ C(b1 ,b2 ,b3 ) |z|L2 (0,T ;L2w (Γ−S )) +|Z |L2 (0,T ;L2 (G )) . F F (13) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The observability estimate for (11) reads |zT |L2 FT (Ω;L2 (G )) ≤ C(b1 ,b2 ,b3 ) |z|L2 (0,T ;L2w (Γ−S )) +|Z |L2 (0,T ;L2 (G )) . F (13) F • Proposition 1: Let (z, Z ) ∈ CF ([0, T ]; L2 (Ω; L2 (G ))) ×L2F (0, T ; L2 (G )) solve the equation (11) with the terminal state zT . Then |z|2L2 2 F (0,T ;Lw (Γ−S )) ≤ e Cr2 E|zT |2L2 (G ) . Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The observability estimate for (11) reads |zT |L2 FT (Ω;L2 (G )) ≤ C(b1 ,b2 ,b3 ) |z|L2 (0,T ;L2w (Γ−S )) +|Z |L2 (0,T ;L2 (G )) . F (13) F • Proposition 1: Let (z, Z ) ∈ CF ([0, T ]; L2 (Ω; L2 (G ))) ×L2F (0, T ; L2 (G )) solve the equation (11) with the terminal state zT . Then |z|2L2 2 F (0,T ;Lw (Γ−S )) ≤ e Cr2 E|zT |2L2 (G ) . • Theorem 7: If T > R, then the inequality (13) holds. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We prove the inequality (13) by a global Carleman estimate. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We prove the inequality (13) by a global Carleman estimate. • Let 0 < c < 1 such that cT > R. Put h T 2 i l = λ |x|2 − c t − and θ = e l . 2 (14) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • We prove the inequality (13) by a global Carleman estimate. • Let 0 < c < 1 such that cT > R. Put h T 2 i l = λ |x|2 − c t − and θ = e l . 2 (14) • Proposition 2: Assume that v is an H 1 (Rn )-valued continuous semi-martingale. Put p = θv . We have the following equality −θ(lt + U · ∇l )p dv + U · ∇vdt 1 1 = − d (lt + U · ∇l )p 2 − U · ∇ (lt + U · ∇l )p 2 2 2 1 + ltt + U · ∇(U · ∇l ) + 2U · ∇lt p 2 2 1 + (lt + U · ∇l )(dp)2 + (lt + U · ∇l )2 p 2 . 2 (15) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Proof of the observability estimate: We apply Proposition 2 to the equation (11) with v = z, integrating (15) on (0, T ) × G , and taking mathematical expectation, then we get that −E Z T 0 ≥ λE Z θ 2 (lt + U · ∇l )z(dz + U · ∇zdt)dxdt G Z (cT − 2U · x)θ 2 (T )z 2 (T )dx G +2cλE Z T 0 Z +2(1 − c)λE Γ−S Z 0 +E Z 0 TZ G U · ν c(T − 2t) − 2U · x θ 2 z 2 dΓ−S dt TZ θ 2 z 2 dxdt G θ 2 (lt +U · ∇l ) (b3 z + Z )2 + 2z 2 dxdt. (16) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla By virtue of that z solves the equation (11), we see Z Z (cT − U · x)θ 2 (T )z 2 (T )dx 2λE S d−1 G +2(1 − c)λE +E Z +E Z T 0 0 ≤ 3E Z TZ Z −2cλE 0 T 0 Z θ 2 z 2 dxdt G θ 2 (lt + U · ∇l )(b3 z + Z )2 dxdt G (17) θ 2 (lt + U · ∇l )2 z 2 dxdt G TZ G Z TZ 0 Z θ 2 (b12 z 2 +b22 Z 2 )dxdt U ·ν c(T−2t)−2U · x θ 2 z 2 dΓ−S dt. Γ−S Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Taking λ1 = 2 X 3 |bi |2L∞ (0,T ;L∞ (G )) + |b3 |4L∞ (0,T ;L∞ (G )) + 1 , F F 2(1 − c) i =1 for any λ ≥ λ1 , it holds that 3E Z 0 T Z G θ 2 (b12 + b34 + b32 )z 2 dxdt ≤ 2(1 − c)λE Z T 0 Z (18) θ 2 z 2 dxdt. G Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Since cT > R, we find that E Z 2 2 θ (T , x)z (T , x)dx ≤ E 0 G −2cλE Z TZ Z TZ 0 θ 2 b22 +2λx −2cλt Z 2 dxdt G U ·ν c(T −2t)−2U ·x θ 2 z 2 dΓ−S dt. Γ−S (19) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Since cT > R, we find that E Z 2 2 θ (T , x)z (T , x)dx ≤ E Z TZ 0 G −2cλE Z TZ 0 2 Γ−S e − 2 cλT E Z G ≤ Ce 1 λR 2 2 G U ·ν c(T −2t)−2U ·x θ 2 z 2 dΓ−S dt. • By the definition of θ, we have e 1 θ 2 b22 +2λx −2cλt Z 2 dxdt Z E ≤θ≤e 1 λR 2 2 (19) . Thus, zT2 dx T 0 − 12 cλT 2 Z 2 Z dxdt + E G Z T 0 Z U · νz dΓ−S dt , 2 Γ−S Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Lack of exact controllability • We only consider a very special case of the system (7), that is, G = (0, 1), a1 = 0 and a2 = 1. The argument for the general case is very similar. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Lack of exact controllability • We only consider a very special case of the system (7), that is, G = (0, 1), a1 = 0 and a2 = 1. The argument for the general case is very similar. • Set ( 1, if t ∈ (1 − 2−2i )T , (1 − 2−2i −1 )T , i = 0, 1, · · · , η(t) = −1, otherwise Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Lack of exact controllability • We only consider a very special case of the system (7), that is, G = (0, 1), a1 = 0 and a2 = 1. The argument for the general case is very similar. • Set ( 1, if t ∈ (1 − 2−2i )T , (1 − 2−2i −1 )T , i = 0, 1, · · · , η(t) = −1, otherwise Z T • Lemma 1: Let ξ = η(t)dW (t). It is impossible to find 0 (̺1 , ̺2 ) ∈ L2F (0, T ; R) × L2F (0, T ; R) and x ∈ R with lim E|̺2 (t) − ̺(T )|2 = 0, t→T such that ξ=x+ Z T ̺1 (t)dt + 0 Z T ̺2 (t)dW (t). 0 (20) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Put △ V = {v : v ∈ L2F (0, T ; L2 (0, 1)), v = 0 in (0, T ) × G0 }. Let ξ be given by Lemma 1. Choose a ψ ∈ C0∞ (G0 ) such that |ψ|L2 (G ) = 1 and set yT = ξψ. We will show that yT cannot be attained for any y0 ∈ R, u ∈ L2F (0, T ; R) and v ∈ V. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Put △ V = {v : v ∈ L2F (0, T ; L2 (0, 1)), v = 0 in (0, T ) × G0 }. Let ξ be given by Lemma 1. Choose a ψ ∈ C0∞ (G0 ) such that |ψ|L2 (G ) = 1 and set yT = ξψ. We will show that yT cannot be attained for any y0 ∈ R, u ∈ L2F (0, T ; R) and v ∈ V. • If there exist a u ∈ L2F (0, T ; R) and a v ∈ V such that y (T ) = yT , then by Itô’s formula, we obtain Z ξ = yT ψdx G = Z = Z G G Z TZ Z TZ y0 ψdx − ψyx dxdt + ψ(y + v )dxdW (t) 0 0 G G Z T Z Z T Z y0 ψdx + ψx ydx dt + ψydx dW (t). 0 G 0 G (21) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • It is clear that both Z ψx ydx and G Z ψydx belong to G L2F (0, T ; R). Further, Z Z 2 lim E ψy (t)dx − ψy (T )dx t→T G G Z 2 = lim E ψ y (t) − y (T ) dx = 0. t→T G These, together with (21), contradict Lemma 1. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla 3. Some recent controllability results for SPDEs and open problems Let us consider the following forward stochastic parabolic equation X (aij yi )j dt = [hα, ∇y i + βy + χG0 f ]dt dy − i ,j +(qy + g ) dW (t) in Q, (22) y =0 on Σ, y (0) = y0 in G , with suitable coefficients α, β, p and q. In (22), the initial state y0 ∈ L2F0 (Ω; L2 (G )). The control variable consists of the pair (f , g ) ∈ L2F (0, T ; L2 (G0 )) × L2F (0, T ; L2 (G )). Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The null controllability of (22) is established by S. Tang and X. Zhang(2009, SICON). From which, they proved that system (22) is null controllable with two controls, i.e., one act on the drift term and the other act on the diffusion term. The control in the diffusion term have to be acted on the whole domain where the solution of system (22) defined!!! Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The null controllability of (22) is established by S. Tang and X. Zhang(2009, SICON). From which, they proved that system (22) is null controllable with two controls, i.e., one act on the drift term and the other act on the diffusion term. The control in the diffusion term have to be acted on the whole domain where the solution of system (22) defined!!! • Is it possible to use one control to drive the solution to zero? Positive answer is given by Lü(2011,JFA) for a special case of system (22). Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla Consider the following controlled stochastic heat equation with one control: n X dy − (aij yxi )xj dt = a(t)ydW + χE χG0 fdt in Q, i ,j=1 (23) y =0 on Σ, y (0) = y0 in G , where a(t) ∈ L∞ F (0, T ), E is a measurable subset in (0, T ) with a positive Lebesgue measure (i.e., m(E ) > 0), χE is the characteristic function of E , y0 ∈ L2 (Ω, F0 , P; L2 (G )), the control 2 2 f belongs to L∞ F (0, T ; L (Ω; L (G ))). Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Theorem 8(Qi Lü, 2011, JFA): System (1) is null controllable at time T , i.e., for each initial datum y0 ∈ L2F0 (Ω; L2 (G )), 2 2 there is a control f ∈ L∞ F (0, T ; L (Ω; L (G ))) such that the solution y of system (1) satisfies y (T ) = 0 in G , P-a.s. Moreover, the control f satisfies the following estimate: |f |2L∞ (0,T ;L2 (Ω;L2 (G ))) ≤ C E|y0 |2L2 (G ) . F (24) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Theorem 8(Qi Lü, 2011, JFA): System (1) is null controllable at time T , i.e., for each initial datum y0 ∈ L2F0 (Ω; L2 (G )), 2 2 there is a control f ∈ L∞ F (0, T ; L (Ω; L (G ))) such that the solution y of system (1) satisfies y (T ) = 0 in G , P-a.s. Moreover, the control f satisfies the following estimate: |f |2L∞ (0,T ;L2 (Ω;L2 (G ))) ≤ C E|y0 |2L2 (G ) . F • (24) For general stochastic heat equations, the answer is unknown. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Consider the following stochastic Schrödinger equation: idy +∆ydt = (a1 ·∇y +a2 y )dt +(a3 y +g )dW (t) y =0 y =u y (0) = y0 in Q, on Σ \ Σ0 , on Σ0 , in G . (25) Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • Consider the following stochastic Schrödinger equation: idy +∆ydt = (a1 ·∇y +a2 y )dt +(a3 y +g )dW (t) y =0 y =u y (0) = y0 in Q, on Σ \ Σ0 , on Σ0 , in G . (25) • Here, the initial state y0 ∈ H −1 (G ), the control u ∈ L2F (0, T ; L2 (Γ0 )), g ∈ L2 (0, T ; H −1 (G )), ak (k = 1, 2, 3) are suitable coefficients. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The solution to (25) is defined in the sense of transposition solution, which is not studied in the literature of stochastic PDEs. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla • The solution to (25) is defined in the sense of transposition solution, which is not studied in the literature of stochastic PDEs. • Theorem 9(Qi Lü, 2013, JDE): System (1) is exactly controllable at any time T > 0. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla The main difficulty of the study of the control and observation problems for SPDEs. • 1. The lack of desired theory, i.e., the well-posedness results of the nonhomogeneous boundary value problems for SPDEs, the propagation of singularities results of the solution for SPDEs, etc. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla The main difficulty of the study of the control and observation problems for SPDEs. • 1. The lack of desired theory, i.e., the well-posedness results of the nonhomogeneous boundary value problems for SPDEs, the propagation of singularities results of the solution for SPDEs, etc. • 2. The stochastic settings lead some useful methods invalid, for example, the lost of the compact embedding for the state spaces, i.e., although L2 (Ω; H01 (G )) ⊂ L2 (Ω; L2 (G )), the embedding is not compact, which violates the compactness -uniqueness argument. Another example is that the irregularity of the solution with respect to the time variable, for example, it is very hard to estimate the time derivative of the solutions, which is an obstacle to follow the method by Fernández Cara-Guerrero-Imanuvilov-Puel to study the null controllability of the stochastic Navier-Stokes equation. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla The main difficulty of the study of the control and observation problems for SPDEs. • 3. Generally speaking, the adjoint system is a backward SPDE, whose solution is constituted by two stochastic process. For example, let us recall the adjoint system of (4) is as follows: ( dz = − A⊤ z + C ⊤ Z dt + ZdW , z(T ) = zT ∈ L2FT (Ω; Rn ). t ∈ [0, T ], (26) When studying the observability problems, people now usually treat Z as a nonhomogeneous term. It is hard to use the fact that Z is a part of the solution. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla This area is full of open problems. In my opinion, some of them are as follows. • The observability estimate for stochastic wave equations under the Geometric Control Conditions. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla This area is full of open problems. In my opinion, some of them are as follows. • The observability estimate for stochastic wave equations under the Geometric Control Conditions. • The null controllability for stochastic heat equations with only one control. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla This area is full of open problems. In my opinion, some of them are as follows. • The observability estimate for stochastic wave equations under the Geometric Control Conditions. • The null controllability for stochastic heat equations with only one control. • The null controllability of stochsatic wave equations, stochastic Schrödinger equations, etc. For this, one need to show that either one control is enough or, as the exact controllability problems, two controls are necessary. Outline 1. A toy model 2. Exact controllability of Stochastic transport equations 3. Some recent controlla . Thank you!