
This paper aims at investigating the convergence and quasi- optimality of an adaptive finite element method for control constrained nonlinear elliptic optimal control problems. We derive a posteriori error estimation for both the control, the state and adjoint state variables under controlling by L2-norms where bubble function is a wonderful tool to deal with the global lower error bound. Then a contraction is proved before the convergence is proposed. Furthermore, we find that if keeping the grids sufficiently mildly graded, we can prove the optimal convergence and the quasi-optimality for the adaptive finite element method. In addition, some numerical results are presented to verify our theoretical analysis.
Citation: Zuliang Lu, Fei Huang, Xiankui Wu, Lin Li, Shang Liu. Convergence and quasi-optimality of L2−norms based an adaptive finite element method for nonlinear optimal control problems[J]. Electronic Research Archive, 2020, 28(4): 1459-1486. doi: 10.3934/era.2020077
[1] |
Zuliang Lu, Fei Huang, Xiankui Wu, Lin Li, Shang Liu .
Convergence and quasi-optimality of |
[2] | Ying Liu, Yanping Chen, Yunqing Huang, Yang Wang . Two-grid method for semiconductor device problem by mixed finite element method and characteristics finite element method. Electronic Research Archive, 2021, 29(1): 1859-1880. doi: 10.3934/era.2020095 |
[3] | Xiu Ye, Shangyou Zhang . A stabilizer free WG method for the Stokes equations with order two superconvergence on polytopal mesh. Electronic Research Archive, 2021, 29(6): 3609-3627. doi: 10.3934/era.2021053 |
[4] |
Mingxia Li, Dongying Hua, Hairong Lian .
On |
[5] | Wenya Qi, Padmanabhan Seshaiyer, Junping Wang . A four-field mixed finite element method for Biot's consolidation problems. Electronic Research Archive, 2021, 29(3): 2517-2532. doi: 10.3934/era.2020127 |
[6] | Liupeng Wang, Yunqing Huang . Error estimates for second-order SAV finite element method to phase field crystal model. Electronic Research Archive, 2021, 29(1): 1735-1752. doi: 10.3934/era.2020089 |
[7] |
Guifen Liu, Wenqiang Zhao .
Regularity of Wong-Zakai approximation for non-autonomous stochastic quasi-linear parabolic equation on |
[8] | Jianguo Huang, Sen Lin . A C0P2 time-stepping virtual element method for linear wave equations on polygonal meshes. Electronic Research Archive, 2020, 28(2): 911-933. doi: 10.3934/era.2020048 |
[9] | Shuhao Cao . A simple virtual element-based flux recovery on quadtree. Electronic Research Archive, 2021, 29(6): 3629-3647. doi: 10.3934/era.2021054 |
[10] | Hyung-Chun Lee . Efficient computations for linear feedback control problems for target velocity matching of Navier-Stokes flows via POD and LSTM-ROM. Electronic Research Archive, 2021, 29(3): 2533-2552. doi: 10.3934/era.2020128 |
This paper aims at investigating the convergence and quasi- optimality of an adaptive finite element method for control constrained nonlinear elliptic optimal control problems. We derive a posteriori error estimation for both the control, the state and adjoint state variables under controlling by L2-norms where bubble function is a wonderful tool to deal with the global lower error bound. Then a contraction is proved before the convergence is proposed. Furthermore, we find that if keeping the grids sufficiently mildly graded, we can prove the optimal convergence and the quasi-optimality for the adaptive finite element method. In addition, some numerical results are presented to verify our theoretical analysis.
Since the pioneer work in adaptive finite element methods was proposed by Babu
When Dörfler [15] presented a marking strategy aiming at electing the set of elements for refinement, based on the error indicators which was controlled by the control, the state and adjoint state, adaptive finite element algorithm was put on the stage of academic research. He provided a fineness assumption on the initial grid
Convergence and quasi-optimality are the two key factors of adaptive finite element methods. It was noteworthy that Mekchay and Nochetto [29] extended the convergence result of Morin, Nochetto and Siebert [31] for general second order linear elliptic partial differential equations by introducing a novel concept that was the total error which was the sum of the energy errors adding the oscillations. This provided a valuable empirical basis for future scholars' work on convergence analysis. Meanwhile, Binev, Dahmen and Devore [4] firstly presented the property of optimality. Later, a large number of scholars participate in the study of the property. For example, Carstensen and Hoppe [6] proposed convergence and quasi-optimality which were established for the Raviart Thomas finite element method. Gong and Yan [20] considered the convergence analysis of adaptive finite element method for elliptic optimal control problems with pointwise control constraints.
As to our best knowledge, nonlinear optimal control problems have gradually penetrated into many fields of scientific research and engineering technology. Chen and Lu [11] investigated adaptive fully-discrete finite element methods for semilinear parabolic quadratic boundary optimal control problems. Gaevskaya, Hoppe, Iliash and Kieweg [17] developed an adaptive finite element method for a class of distributed optimal control problems with control constraints, and found requirement
Leng and Chen [23] proved the convergence and the quasi-optimality of an adaptive element method with integral control constraints while we extend the result of [23] to a nonlinear optimal control problem with integral control constraint on
Here are some notations will be used in this paper. Let
The rest of our paper is arranged as follows. In Section 2, we give what the optimal control problems we want to investigate and some basic notations must be used. Then the a posteriori error estimation is obtained and an adaptive algorithm is proposed in Section 3. In Section 4, we use quasi-orthogonality and discrete local upper bound to prove the convergence of the adaptive finite element method and so is the quasi-optimality for details in Section 5. In the end, some numerical simulations is given to verify our theoretical analysis.
In this section we first introduce some basic notations, and then we show what the nonlinear optimal control problem we discussed about.
Lemma 2.1. [21] For some constants
chT≤hTh|T≤ChT, | (1) |
||∇hTh||∞≤μ, | (2) |
where all grids satisfied above are denoted by
In this paper we mainly enter into meaningful discussions with the following nonlinear optimal control problem:
minu∈Uad{12‖y−yd‖20+α2‖u‖20},−Δy+ϕ(y)=f+u,in Ω, | (3) |
y=0,on ∂Ω, | (4) |
where
a(y,v)+(ϕ(y),v)=(f+u,v),∀ v∈V, |
where
a(y,v)=∫Ω∇y⋅∇vdxand|||v|||=√a(v,v). |
Then the nonlinear optimal control problem can be restated as follows
minu∈Uad{12‖y−yd‖20+α2‖u‖20}, | (5) |
a(y,v)+(ϕ(y),v)=(f+u,v),∀ v∈V. | (6) |
It is well known [26,27] that the nonlinear optimal control problem has at least one solution
a(y,v)+(ϕ(y),v)=(f+u,v),∀ v∈V, | (7) |
a(q,p)+(ϕ′(y)p,q)=(y−yd,q),∀ q∈V, | (8) |
(αu+p,v−u)≥0,∀ v∈Uad. | (9) |
Since the coercivity of
minuh∈Uhad{12‖yh−yd‖20+α2‖uh‖20}, | (10) |
a(yh,v)+(ϕ(yh),v)=(f+uh,v),∀ v∈Vh. | (11) |
Similarly the nonlinear optimal control problem (10)-(11) has at least one solution
a(yh,v)+(ϕ(yh),v)=(f+uh,v),∀ v∈Vh, | (12) |
a(q,ph)+(ϕ′(yh)ph,q)=(yh−yd,q),∀ q∈Vh, | (13) |
(auh+ph,vh−uh)≥0,∀ vh∈Uhad. | (14) |
Based on [12,14,21,27], we have the following Lemmas in order to derive a
Lemma 2.2. [21] Suppose that
||y||2≤C||f+u||0, |
and apparently the assumption is valid for
Lemma 2.3. [12] Assume that
ϕ(y)−ϕ(p)=−˜ϕ′(y)(p−y)=−ϕ′(p−y)+˜ϕ″(y)(p−y), |
for any
˜ϕ′(y)=∫10ϕ′(y+s(p−y))ds,˜ϕ″(y)=∫10(1−s)ϕ″(p+s(p−y))ds, |
are bounded functions in
Lemma 2.4. [14,27] For all
||∇v||0,∂T∖∂Ω≤Ch−1/2T||∇v||0+Ch1/2T|v|2. | (15) |
In this section, we will recall a residual-based a posteriori error estimation for nonlinear elliptic equations. For the model problem that we studied in Section 2, a reliable and efficient a posteriori estimation will be obtained. In the end of this section, an adaptive finite element algorithm will be introduced.
Here we define some error indicators.
η21,Th(ph,T)=h2T‖∇ph‖20,T,η22,Th(uh,yh,T)=h4T‖f+uh−ϕ(yh)‖20,T+h3T‖(∇yh)⋅n‖20,∂T∖∂Ω,η23,Th(yh,ph,T)=h4T‖yh−yd−ϕ′(yh)ph‖20,T+h3T‖(∇ph)⋅n‖20,∂T∖∂Ω,osc2Th(f,T)=h4T‖f−fT‖20,T,osc2Th(yh−yd,T)=h4T‖(yh−yd)−(yh−yd)T‖20,T, |
where
η21,Th(ph,ω)=∑T∈ωη21,Th(ph,T),osc2Th(f,ω)=∑T∈ωosc2Th(f,T). |
Similarly, we have
Lemma 3.1. Let
||y−yh||0≤C(||u−uh||0+|||hTh(y−yh)|||),||p−ph||0≤C(||u−uh||0+|||hTh(y−yh)|||+|||hTh(p−ph)|||), |
for sufficiently small
Proof. Suppose that
a(yh,v)+(ϕ(yh),v)=(f+uh,v),∀ v∈V, | (16) |
a(q,ph)+(ϕ′(yh)ph,q)=(yh−yd,q),∀ q∈V. | (17) |
Employing the Galerkin orthogonality, the approximation properties, Lemma 2.2, and
||yh−yh||0≤C|||hTh(yh−yh)|||,||ph−ph||0≤C|||hTh(ph−ph)|||. |
It has been proved
|||y−yh|||≤C||u−uh||0. | (18) |
|||p−ph|||≤C(||u−uh||0+||y−yh||0). | (19) |
By using the triangle inequality, we have
||y−yh||0≤||y−yh||0+||yh−yh||0,||p−ph||0≤||p−ph||0+||ph−ph||0. |
In connection with what we discussed above, the triangle inequality and Lemma 2.1, it is easy to prove the prevenient results in Lemma 3.1.
Now we are in the position to derive a posteriori error estimation for both the control, the state and adjoint state variables.
Theorem 3.2. Let
||u−uh||20+||y−yh||20+||p−ph||20≤c(η21,Th(ph,Th)+η22,Th(uh,yh,Th)+η23,Th(yh,ph,Th)), |
and apparently a global lower bound
C(η21,Th(ph,Th)+η22,Th(uh,yh,Th)+η23,Th(yh,ph,Th))≤||u−uh||20+||y−yh||20+||p−ph||20+osc2Th(f,Th)+osc2Th(yh−yd,Th), |
where
Proof. In view of Lemma 7.3.1 in [27], we similarly derive that
||u−uh||20≤Cη21,Th(ph,Th)+C||ph−ph||20, | (20) |
where
Let
||ph−ph||20=a(∇ξ,∇(ph−ph))+(ϕ′(yh)(ph−ph),ξ)=(∇(ξ−ξI),∇(ph−ph))+(ϕ′(yh)ph−ϕ′(yh)ph,ξ−ξI)+(∇ξI,∇(ph−ph))+(ϕ′(yh)ph−ϕ′(yh)ph,ξI)+((ϕ′(yh)−ϕ′(yh))ph,ξ)=∑T∈Th∫T(yh−yd−ϕ′(yh)ph)(ξ−ξI)−∑T∈Th∫∂T[(∇ph)⋅n](ξ−ξI)ds+(yh−yh,ξI)+((ϕ′(yh)−ϕ′(yh))ph,ξ) |
≤C∑T∈Thh2T||yh−yd−ϕ′(yh)ph||0,T|ξ|2,T+C∑∂T∖∂Ωh3/2T(∫T[(∇ph)⋅n]2)1/2||ξ||2,∂T∖∂Ω+||yh−yh||0||ξ||0+C||ϕ′(yh)−ϕ′(yh)||0||ph||0||ξ||0,∞≤C∑T∈Thh4T∫T(yh−yd−ϕ′(yh)ph)2+C∑∂T∖∂Ωh3T[(∇ph)⋅n]2+||yh−yh||20+Cδ||ξ||22, |
where
||ph−ph||20≤Cη23,Th(yh,ph,Th)+C||yh−yh||20. | (21) |
Similarly there is going to be proved by letting
||yh−yh||20=a(∇(yh−yh),∇˜ξ)+(ϕ(yh)−ϕ(yh),˜ξ)=(∇(yh−yh),∇(˜ξ−˜ξI))+(ϕ(yh)−ϕ(yh),˜ξ−˜ξI)=∑T∈Th∫T(f+uh−ϕ(yh))(˜ξ−˜ξI)−∑∂T∖∂Ω∫T[(∇yh)⋅n](˜ξ−˜ξI)≤C∑T∈Thh4T∫T(f+uh−ϕ(yh))2+C∑∂T∖∂Ωh3T∫T[(∇yh)⋅n]2+Cδ||˜ξ||20, |
where
||yh−yh||20≤Cη22,Th(uh,yh,Th). | (22) |
It is easy to derive the expect upper bound by combining with (20)-(22).
Next we are going to deduce the global lower error bound through the standard bubble function [1,19]. Similar to Lemma 3.7 in [19], it can be similarly proved that there exists polynomial
∫Th4T((yh−yd)T−ϕ′(yh)ph)2=∫Th4T((yh−yd)T−ϕ′(yh)ph)wT, | (23) |
and apparently
||wT||20,T≤C∫T((yh−yd)T−ϕ′(yh)ph)2, | (24) |
ch−2T||wT||20,T≤|wT|22,T≤Ch−2T||wT||20,T. | (25) |
Then it follows from (23) and (24) that
∫Th4T((yh−yd)T−ϕ′(yh)ph)2=∫Th4T((yh−yd)T−ϕ′(yh)ph)wT,=∫Th4T(yh−yd−ϕ′(yh)ph)wT+∫Th4T((yh−yd)T−(yh−yd))wT≡I1+I2. |
Combining with (2.8) and Lemma 2.4, there holds
I1=∫Th4T(ph−p)ΔwT+∫Th4T(yh−y)wT+∫Th4Tϕ′(y)(p−ph)phwT+∫Th4T((ϕ′(y)−ϕ′(yh))phwT≤C(||p−ph||20,T+||yh−y||20,T+h4T||ϕ′(y)||0,T||p−ph||0,T||ph||0,T||wT||0,∞,T+h4T||ϕ′(y)−ϕ′(yh)||0,T||ph||0,T||wT||0,∞,T)+C(h2T||wT||20,T+h4T|wT|22,T)≤C(||p−ph||20,T+||yh−y||20,T)+Cδh4T||wT||20, | (26) |
where
Similarly, we have
I2≤C∫Th4T((yh−yd)T−(yh−yd)2+Cδh4T||wT||0,T. | (27) |
Hence by using the Cauchy inequality with the help of (25)-(27), we obtain
∫Th4T((yh−yd)T−ϕ′(yh)ph)2≤C(||p−ph||20,T+||yh−y||20,T+∫Th4T((yh−yd)T−(yh−yd))2)+Ch4T∫Th4T((yh−yd)T−ϕ′(yh)ph)2. |
Then it brings about
∫Th4T((yh−yd)−ϕ′(yh)ph)2≤C(∫Th4T((yh−yd)T−ϕ′(yh)ph)2+∫Th4T((yh−yd)T−(yh−yd))2)≤C(||p−ph||20,T+||yh−y||20,T+∫Th4T((yh−yd)T−(yh−yd))2). | (28) |
Then we need to use the new bubble functions defined in [19] to both in deal with the jump. Similar to [18,19], it can be similarly proved that there exists polynomial
∫∂Th3T[(∇ph)⋅n]2=∫∂Th3T[(∇ph)⋅n]w∂T, | (29) |
and apparently
||w∂T||20,∂T∖∂Ω≤C∫∂Th3T[(∇ph)⋅n]2, | (30) |
ch−2T||w∂T||20,∂T∖∂Ω≤||w∂T||22,∂T∖∂Ω≤Ch−2T||w∂T||20,∂T∖∂Ω. | (31) |
And then it follows from (29) and(30) that
∫∂Th3T[(∇ph)⋅n]2=∫∂Th3T[(∇ph)⋅n]w∂T=∫∂T[(∇ph)⋅n−(∇p)⋅n]w∂T=∫∂Th3T(p−ph)Δw∂T+∫∂Th3T(y−yd−ϕ′(y)p)w∂T≡I3+I4. |
Similarly, it can be deduced that
I3≤Cδh3T||w∂T||20,∂T∖∂Ω+C||p−ph||20,∂T∖∂Ω. |
Combining with (13) and Lemma 2.4, there holds
I4=∫∂Th3T(yh−yd−ϕ′(yh)ph)w∂T+∫∂Th3T(y−yh)w∂T+∫∂Th3Tϕ′(y)(p−ph)w∂T+∫∂Th3T˜ϕ″(y)(y−yh)phw∂T≤C(||y−yh||20,∂T∖∂Ω+∫∂Th3T(yh−yd−ϕ′(yh)ph)2+h3T||ϕ′(y)||0,∂T∖∂Ω||p−ph||0,∂T∖∂Ω||w∂T||0,∞,∂T∖∂Ω+h3T||˜ϕ″(y)||0,∂T∖∂Ω||y−yh||0,∂T∖∂Ω||ph||0,∂T∖∂Ω||w∂T||0,∞,∂T∖∂Ω)+Cδ(h3T||w∂T||20,∂T∖∂Ω+h3T|w∂T|22,∂T∖∂Ω)≤C(||y−yh||20,∂T∖∂Ω+||p−ph||20,∂T∖∂Ω+∫∂Th3T(yh−yd−ϕ′(yh)ph)2)+Cδh3T||w∂T||20,∂T∖∂Ω. |
Hence by using the Cauchy inequality with the help of (29)-(30), we have
∑∂T∖∂Ω∫∂Th3T[(∇ph)⋅n]2≤C(||p−ph||20+||y−yh||20)+C∑T∈Th∫Th4T(yh−yd−ϕ′(yh)ph)2, | (32) |
where
In connection with (28) and (32), it is easy to get that
η23,Th(yh,ph,Th)=∑T∈Th∫Th4T(yh−yd−ϕ′(yh)ph)2+∑∂T∖∂Ω∫∂Th3T([∇ph]⋅n)2≤C‖ph−p‖20+C‖y−yh‖20+C∑T∈Th∫Th4T((yh−yd)−(yh−yd)T)2=C‖ph−p‖20+C‖y−yh‖20+Cosc2Th(yh−yd,Th). |
It can also be deduced that
η22,Th(uh,yh,Th)=∑T∈Th∫Th4T(f+uh−ϕ(yh))2+∑∂T∖∂Ω∫∂Th3T([∇yh]⋅n)2≤C‖yh−y‖20+C‖u−uh‖20+C∑T∈Th∫Th4T(f−fT)2=C‖yh−y‖20+C‖u−uh‖20+Cosc2Th(f,Th). |
Above-mentioned results tell the proof of Theorem 3.2 is accomplished.
Theorem 3.2 gives a reliable and efficient posteriori error estimations for the sum of the
Algorithm 3.1. Adaptive finite element algorithm for nonlinear optimal control problems:
(o) Given an initial mesh
(1) Solve the discrete nonlinear optimal control problem (12)-(14), then obtain approximate solution
(2) Compute the local error estimator
(3) Select a minimal subset
η2Thk(Mhk)≥θη2Thk(Thk), |
where
(4) Refine
(5) Solve the discrete nonlinear optimal control problem (12)-(14), then obtain approximate solution
(6) Set
In this section, we will do our best to demonstrate the convergence while we first give some properties which take vital significance to the proof of the convergence and even the quasi-optimality for the error indicators and the data oscillations before we begin to show the convergence analysis.
Lemma 4.1. Let
||u−uh||20+|||hTh(y−yh)|||2+|||hTh(p−ph)|||2≤C(η21,Th(ph,Th)+η22,Th(uh,yh,Th)+η23,Th(yh,ph,Th)), |
where
Proof. By applying (18), the triangle inequality, and Lemma 2.1, we obtain
|||hTh(y−yh)|||2≤|||hTh(y−yh)|||2+|||hTh(yh−yh)|||2≤(||∇hTh||∞|||y−yh|||)2+|||hTh(yh−yh)|||2≤C||u−uh||20+|||hTh(yh−yh)|||2. | (33) |
Similar to the proof of Lemma 3.3 in [21], we deduce that
|||hTh(yh−yh)|||2≤η22,Th(uh,yh,Th). | (34) |
Analogously, the following conclusions can be drawn
|||hTh(p−ph)|||2≤C||u−uh||20+C||y−yh||2+|||hTh(ph−ph)|||2, | (35) |
and apparently
|||hTh(ph−ph)|||2≤η23,Th(yh,ph,Th). | (36) |
It is obvious to get the expected result in Lemma 4.1 via using (33)-(36), and the upper bound in Theorem 3.2.
Next, we gives a stability result for error indicators which is can be found in Lemma 3.4 in [21], Lemma 4.1 in [23], and even Proposition 3.3 in [8], and so on.
Lemma 4.2. For
η1,Th(ph1,T)−η1,Th(ph2,T)≤C(|||hTh(ph1−ph2)|||T+||∇hTh||∞,T||ph1−ph2||0,T), | (37) |
η2,Th(uh1,yh1,T)−η2,Th(uh2,yh2,T)≤C(h2T||uh1−uh2||0,T+|||hTh(yh1−yh2)|||ωT+||∇hTh||∞,T||yh1−yh2||0,ωT), | (38) |
η3,Th(yh1,ph1,T)−η3,Th(yh2,ph2,T)≤C(h2T||yh1−yh2||0,T+|||hTh(ph1−ph2)|||ωT+||∇hTh||∞,T||ph1−ph2||0,ωT), | (39) |
oscTh(yh1−yd,T)−oscTh(yh2−yd,T)≤C(|||hTh(yh1−yh2)|||T+||∇hTh||∞,T||yh1−yh2||0,T), | (40) |
where
Proof. We first prove (37) while (40) can be just proved similarly. Consulting the literatures [1,21,25], namely the trace inequality, there exists
||v||0,∂T≤C(h−1/2T||v||0,T+h1/2T||v||1,T), | (41) |
for arbitrary
||[∇(ph1−ph2)]⋅n||0,∂T∖∂Ω≤Ch−1/2T|||ph1−ph2|||ωT. | (42) |
Recalling (1) in Lemma 1, we know that
hT|||ph1−ph2|||ωT≤C||hTh∇(ph1−ph2)||0,ωT≤C(|||hTh∇(ph1−ph2)|||ωT+||(ph1−ph2)∇hTh||0,ωT). | (43) |
Recalling the definition of
η1,Th(ph1,T)≤η1,Th(ph2,T)+h3/2T||[∇(ph1−ph2)]⋅n||0,∂T∖∂Ω. | (44) |
Then it is easy to derive the desired result (37) by adopting (42)-(44).
Next we are to prove (38) while (39) can be proved similarly. We calculate while applying the inequality (15) in Lemma 2.4 for the edge
h3/2T||∇(yh1−yh2)||0,T∩T′≤h3/2T(||∇(yh1−yh2)T||0,T∩T′+||∇(yh1−yh2)T′||0,T∩T′)≤Ch3/2T(h−1/2T||∇(yh1−yh2)||0,T+h1/2T|∇(yh1−yh2)|0,T+h−1/2T||∇(yh1−yh2)||0,T′+h1/2T|∇(yh1−yh2)|0,T′)≤C(|||hTh(yh1−yh2)|||0,T∪T′+||(yh1−yh2)∇hTh||0,T∪T′. | (45) |
Recalling the definition
η2,Th(uh1,yh1,T)≤η2,Th(uh2,yh2,T)+(h3T||[∇(yh1−yh2)]||0,∂T∖∂Ω+h4T||ϕ(yh1)−ϕ(yh2)||20,T)12≤η2,Th(uh2,yh2,T)+(h4T||Δ(yh1−yh2)||20,T+h3T||[∇(yh1−yh2)]||0,∂T∖∂Ω+h4T||ϕ′(yh1)||0,T||yh1−yh2||20,T)12≤η2,Th(uh2,yh2,T)+(h3T||[∇(yh1−yh2)]||0,∂T∖∂Ω+h4T||uh1−uh2||20,T+h4T|||yh1−yh2|||T)12. | (46) |
Then it is easy to deduce the expected result (38) by connecting (45) into (46).
Lemma 4.3. For
η21,˜Th(˜p,˜Th)−(1+σ)[η21,Th(p,Th)−λη21,Th(p,Mh)]≤C(1+1σ)(∑T∈˜Thh4T||uh−˜uh||20,T+|||h˜Th(yh−˜yh)|||), | (47) |
η22,˜Th(˜uh,˜yh,˜Th)−(1+σ)[η22,Th(uh,yh,Th)−λη22,Th(uh,yh,Mh)]≤C(1+1σ)(∑T∈˜Thh4T||uh−˜uh||20,T+|||h˜Th(yh−˜yh)|||2+||∇h˜Th||2∞||yh−˜yh||20), | (48) |
η23,˜Th(˜yh,˜ph,˜Th)−(1+σ)[η23,Th(yh,ph,Th)−λη23,Th(yh,ph,Mh)]≤C(1+1σ)(∑T∈˜Thh4T||uh−˜uh||20,T+|||h˜Th(ph−˜ph)|||2+||∇h˜Th||2∞||ph−˜ph||20), | (49) |
osc2Th(yh−yd,Th∩˜Th)−2osc2˜Th(˜yh−yd,Th∩˜Th)≤2C(|||hTh(yh−˜yh)|||2+||∇h˜Th||2∞||yh−˜yh||20), | (50) |
where
Proof. We just prove (49) and (50). The proofs of (47) and (48) are similar with (49). Employing the Young's inequality with parameter
η23,˜Th(˜yh,˜ph,~Th)−η23,˜Th(yh,ph,~Th)≤C(1+1σ)(∑T∈Thh4T‖uh−˜uh‖20,T+|||h˜Th(ph−˜ph|||2+||∇h˜Th||2∞||ph−˜ph||20)+δη23,~Th(yh,ph,~Th). | (51) |
For a marked element
∑T∈˜ThT′η23,˜Th(y,p,T)≤2−3b2η23,Th(y,p,T′). |
For
η23,˜Th(y,p,T′)≤η23,Th(y,p,T′), |
which can use to derive that
η23,˜Th≤2−3b2η23,Th(y,p,Mh)+η22,Th(y,p,Th∖Mh)=η23,Th(y,p,Th)−λη23,Th(y,p,Mh). | (52) |
Then adding (52) into (51) and rearranging the terms can obtain the expected result (49).
Next, for arbitrary
osc2Th(yh−yd,T)−2osc2Th(˜yh−yd,T)≤2C(|||hTh(yh−˜yh)|||T+||∇hTh||∞,T||yh−˜yh||0,T), | (53) |
where the Young's inequality have been applied and
In order to facilitate computation, we introduce the following new notation
e2hk=|||hThk(y−yhk)|||2+|||hThk(p−phk)|||2,E2hk=|||hThk+1(yhk+1−yhk)|||2+|||hThk+1(phk+1−phk)|||2,η2Thk(ω)=η21,Thk(phk,ω)+η22,Thk(uhk,yhk,ω)+η23,Thk(yhk,phk,ω),osc2Thk(ω)=osc2Thk(f,ω)+osc2Thk(yhk−yd,ω), |
for
|||hThk+1(v−vhk+1)|||=|||hThk+1(v−vhk)|||−|||hThk+1(vhk+1−vhk)|||−2a(hThk+1(v−vhk+1),hThk+1(vhk+1−vhk)), | (54) |
so that we obtain the quasi-orthogonality in Lemma 4.4 by estimating the last term of (54).
Lemma 4.4. For any
(1−ϵ)|||hThk+1(y−yhk+1)|||2−(1+ϵ)|||hThk(y−yhk)|||2+|||hThk+1(yhk+1−yhk)|||2 |
≤Cϵ(μ2+h20)(||uh−uhk+1||20+||yh−yhk||20+||y−yhk+1||20), | (55) |
(1−ϵ)|||hThk+1(p−phk+1)|||2−(1+ϵ)|||hThk(p−phk)|||2+|||hThk+1(yhk+1−yhk)|||2≤Cϵ(μ2+h20)(||uh−uhk+1||20+||ph−phk||20+||p−phk+1||20), | (56) |
where
Proof. We just prove (55) while (56) can be proved in a similar way. Let
a(hThk+1(y−yhk+1),hThk+1(yhk+1−yhk))=a(hThk+1(y−yhk+1),hThk+1(yhk+1−yhk))+a(hThk+1(yhk+1−yhk+1),hThk+1(yhk+1−yhk)). | (57) |
Similar to the proof of Lemma 4.3 in [23], we can estimate the second term of the right side of (57) as shown below
a(hThk+1(yhk+1−yhk+1),hThk+1(yhk+1−yhk))≤ϵ8(|||hThk+1(yhk+1−yhk+1)|||2+|||hThk+1(yhk+1−yhk|||2)+C(1+1ϵ)μ2(||yhk+1−yhk+1||20+||yhk+1−yhk||20). |
Next we will subdivide the proof. By applying (2), (18), and the triangle inequality, we have
|||hThk+1(yhk+1−yhk+1)|||≤|||hThk+1(y−yhk+1)|||+|||hThk+1(y−yhk+1)|||≤μ||y−yhk+1||0+|||hThk+1∇(y−yhk+1)|||+|||hThk+1(y−yhk+1)|||≤C(μ+h0)||u−uhk+1||0+|||hThk+1(y−yhk+1)|||, |
in which we use the same way to derive that
|||hThk+1(yhk+1−yhk)|||≤C(μ+h0)||u−uhk+1||0+|||hThk+1(y−yhk)|||. |
It follows from the triangle inequality that
||yhk+1−yhk+1||0≤C(||u−uhk+1||0+||y−yhk+1||0), |
and apparently
||yhk+1−yhk||0≤C(||u−uhk+1||0+||y−yhk||0). |
Then in connection with what we discuss above to deduce that
a(hThk+1(yhk+1−yhk+1),hThk+1(yhk+1−yhk))≤ϵ4(|||hThk+1(y−yhk+1)|||2+|||hThk+1(y−yhk)|||2)+Cϵ(μ2+h20)(||u−uhk+1||20+||y−yhk||20+|||y−yhk|||2). | (58) |
For the first term of the right side of (57), we obtain
a(hThk+1(y−yhk+1),hThk+1(yhk+1−yhk))=((y−yhk+1)∇hThk+1,∇(hThk+1(yhk+1−yhk)))+(hThk+1∇(y−yhk+1),∇(hThk+1(yhk+1−yhk)))≤ϵ4(|||hThk+1(y−yhk+1)|||2+|||hThk+1(y−yhk)|||2)+Cϵ(μ2+h20)||u−uhk+1||20. | (59) |
Contacting (57), (58) and (59) to gain
a(hThk+1(y−yhk+1),hThk+1(yhk+1−yhk))≤Cϵ(μ2+h20)(||u−uhk+1||20+||y−yhk||20+|||y−yhk|||2)+ϵ2(|||hThk+1(y−yhk+1)|||2+|||hThk+1(y−yhk)|||2). | (60) |
Because of
|||hThk+1(y−yhk)|||≤|||hThk+1∇(y−yhk)|||+μ||y−yhk||0≤|||hThk(y−yhk)|||+2μ||y−yhk||0, |
such that
|||hThk+1(y−yhk)|||2≤(1+ϵ)|||hThk(y−yhk)|||2+4μ2(1+1ϵ)||y−yhk||20. | (61) |
Then we can obtain (55) by combing with (54), (60), and (61).
Theorem 4.5. Let
e2hk+1+γη2hk+1(Thk+1)≤α(e2hk+γη2hk(Thk)), | (62) |
apparently providing
Proof. Taking Lemma 3.1, Lemma 4.3, and Lemma 4.4 into account, we can easy to derive that
e2hk≤Cη2Thk(Thk), | (63) |
(1−ϵ)e2hk+1≤(1+ϵ)2e2hk−E2hk+Cϵ(μ2+h20)(||u−uhk+1||20+||y−yhk||20+||y−yhk+1||20+||p−phk||20+||p−phk+1||20), | (64) |
η2Thk+1(Thk+1)≤(1+σ)(η2Thk(Thk)−λη2Thk(Mhk))+C(1+1σ)[(μ2+h20)(||u−uhk||20+||u−uhk+1||20+||y−yhk||20+||y−yhk+1||20+||p−phk||20+||p−phk+1||20)+E2hk]. | (65) |
Simplifying (64) and (65) by employing Lemma 3.1 and Lemma 4.1, we have
(1−ϵ)e2hk+1≤(1+ϵ)2e2hk−E2hk+Cϵ(μ2+h20)(η2Thk(Thk)+η2Thk+1(Thk+1)), | (66) |
η2Thk+1(Thk+1)≤(1+σ)(η2Thk(Thk)−λη2Thk(Mhk))+C(1+1σ)[(μ2+h20)(η2Thk(Thk)+η2Thk+1(Thk+1))+E2hk]. | (67) |
Multiplying (67) with
(1−ϵ)e2hk+1+γ1η2Thk+1(Thk+1)≤(1+ϵ)2e2hk+γ1(1+σ)(η2Thk(Thk)−λη2Thk(Mhk))+γ2(μ2+h20)(η2Thk(Thk)+η2Thk+1(Thk+1)), |
where
(1−ϵ)e2hk+1+[γ1−γ2(μ2+h20)]η2Thk+1(Thk+1)≤((1+ϵ)2−γ3γ1(1+σ)λθC)e2hk+[γ1(1+σ)(1−(1−γ3)λθ)+γ2(μ2+h20)]η2Thk(Thk), |
where
γ=γ1−γ2(μ2+h20)1−ϵ,α1=1−1Cγ3γ1(1+σ)λθ−ϵ(3+ϵ)1−ϵ,α2=γ1(1+σ)(1−(1−γ3)λθ)+γ2(μ2+h20)γ1−γ2(μ2+h20), |
where
e2hk+1+γη2Thk+1(Thk+1)≤α1e2hk+α2γη2Thk(Thk), |
which tells a contraction property, namely (62), in Algorithm 3.1 if selecting
Theorem 4.6. Let
‖u−uhk‖20+‖y−yhk‖20+‖p−phk‖20→0ask→∞. |
Proof. It follows from Lemma 4.1 and Theorem 4.5 that we can obviously get
||u−uhk||0→0ask→∞. |
Then combining with Lemma 3.1, it is distinct to get the desired result in Theorem 4.6.
In this section, we consider the quasi-optimality for the adaptive finite element method. Firstly we give the notations interpretation. For any
#(Th⊕˜Th)≤#Th+#˜Th−#Th0. | (68) |
According to [14,21,23,24], we need to define a function approximation class
As:={(u,y,p,yd)∈L2(Ω)×H10(Ω)×H10(Ω)×L2(Ω):|(u,y,p,yd)|s<+∞}, |
where
|(u,y,p,yd)|s:=supN>0NsinfTh∈TNinf(uh,yh,ph)∈Uhad×Vh×Vh{‖u−uh‖20+‖y−yh‖20+‖p−ph‖20+osc2Th(Th)}12, |
and
TN:={Th∈T:#Th−#Th0≤N0}. |
We need a local upper bound for the distance between two nested solutions consulting [8] in order to illustrate the quasi-optimality of an adaptive finite element method due to the errors here can only be estimated by using refined element indices without buffer.
Lemma 5.1. Let
||uh−˜uh||20+||yh−˜yh||20+||ph−˜ph||20≤Cη2Th(Rh), | (69) |
where
Proof. According to the optimal condition (14), we obtain
(αuh+ph,v−uh)≥0,∀ v∈Vh,(α˜uh+˜ph,v−˜uh)≥0,∀ v∈V˜h, |
thus getting
α||uh−˜uh||20=(αuh,uh−˜uh)−(α˜uh,uh−˜uh)≤(˜ph−ph,uh−˜uh)+(αuh+ph,uh−˜uh). | (70) |
For
(˜ph−ph,uh−˜uh)=(S∗˜Th(S˜Th(f+˜uh)−yd)−S∗Th(STh(f+uh)−yd),uh−˜uh)=(S∗˜Th(S˜Th(f+˜uh)−yd)−S∗˜Th(S˜Th(f+uh)−yd),uh−˜uh)+(S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd),uh−˜uh)=(S∗˜Th(S˜Th(˜uh−uh)),uh−˜uh)+(S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd),uh−˜uh)=(S˜Th(˜uh−uh),S˜Th(uh−˜uh))+(S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd),uh−˜uh)≤(S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd),uh−˜uh). |
In order to estimate the right side of the above inequality, we suppose that
a(φ,v)=(S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd),v),∀ v∈H10(Ω). |
By applying the duality arguments, we can gain that
||S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd)||20=a(φ,S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd))=a(φ−ξThφ,S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd))+(S˜Th(f+uh)−STh(f+uh),ξThφ−φ)+(S˜Th(f+uh)−STh(f+uh),φ), | (71) |
in which
a(φ−ξThφ,S∗˜Th(S˜Th(f+uh)−yd)−S∗Th(STh(f+uh)−yd))=a(φ−ξThφ,S∗˜Th(S˜Th(f+uh)−yd)−S∗˜Th(STh(f+uh)−yd))+a(φ−ξThφ,S∗˜Th(STh(f+uh)−yd)−S∗Th(STh(f+uh)−yd))≤C(||φ||2||S˜Th(f+uh)−STh(f+uh)||0+||h−1Th∇(φ−ξThφ)||0||hTh∇(S∗˜Th((STh(f+uh)−yd)−S∗Th(STh(f+uh)−yd)))||0)≤C||φ||2(||S˜Th(f+uh)−STh(f+uh)||0+||hTh∇(S∗˜Th((STh(f+uh)−yd)−S∗Th(STh(f+uh)−yd)))||0). | (72) |
Similar to Lemma 2 in [14], we infer that
||S˜Th(f+uh)−STh(f+uh)||0≤Cη2,Th(uh,yh,Rh), | (73) |
||hTh∇(S∗˜Th((STh(f+uh)−yd)−S∗Th(STh(f+uh)−yd)))||0≤Cη3,Th(yh,ph,Rh). | (74) |
By using the similar way, we deduce that
(S˜Th(f+uh)−STh(f+uh),ξThφ−φ)≤Cη2,Th(uh,yh,Rh), | (75) |
(S˜Th(f+uh)−STh(f+uh),φ)≤Cη2,Th(uh,yh,Rh). | (76) |
Combining (72)-(76) and Lemma 2.2 with
(˜ph−ph,uh−˜uh)≤C(η2,Th(uh,yh,Rh)+η3,Th(yh,ph,Rh)). | (77) |
Next we are going to estimate the second term on the right side of (70). Assume that
(αuh+ph,uh−˜uh)≤(αuh+ph,vh−˜uh),∀vh∈Uhad. |
Then we set
(αuh+ph,vh−˜uh)T=((ITh−πTh)(αuh+ph),(πTh−ITh)(vh−˜uh)), |
for arbitrary
(αuh+ph,vh−˜uh)≤Cη1,Th(ph,Rh)||uh−˜uh||0. | (78) |
In connection with (70), (77) and (78), we infer that
||uh−˜uh||0≤C(η1,Th(ph,Rh)+η2,Th(uh,yh,Rh)+η3,Th(yh,ph,Rh)). | (79) |
Employing (73)-(76) and the triangle inequality, we deduce that
||ph−˜ph||0=||S∗Th(STh(f+uh)−yd)−S∗˜Th(S˜Th(f+˜uh)−yd)||0≤||S∗Th(STh(f+uh)−yd)−S∗˜Th(S˜Th(f+uh)−yd)||0+||S∗˜Th(S˜Th(f+uh)−yd)−S∗˜Th(S˜Th(f+˜uh)−yd)||0≤C(||uh−˜uh||0+η2,Th(uh,yh,Rh)+η3,Th(yh,ph,Rh)). | (80) |
It is similar to Lemma 2 in [14] that we infer that
||yh−˜yh||0=||STh(f+uh)−S˜Th(f+˜uh)||0≤||STh(f+uh)−S˜Th(f+uh)||0+||S˜Th(f+uh)−S˜Th(f+˜uh)||0≤C(||uh−˜uh||0+η2,Th(uh,yh,Rh)). | (81) |
To sum up, the proof is finished by adopting (79)-(81).
Next lemma tells the Dörfler property on the set
Lemma 5.2. We assume that the marking parameter
e2Thk+osc2Thk(Thk)≤δ(e2Th+osc2Th(Th)), | (82) |
where
η2Th(Rh)≥θη2Th(Th). |
Proof. Combining with (82) and the upper bound in Theorem 4.5, we can obtain
C(1−2δ)η2Th(Th)≤(1−2δ)(e2Th+osc2Th(Th)≤e2Th−2e2Thk+osc2Th(Th)−2osc2Thk(Thk). |
Employing the triangle inequality and the Young's inequality, here holds
||u−uh||20≤2(||u−uhk||20+||uh−uhk||20),||y−yh||20≤2(||y−yhk||20+||yh−yhk||20),||p−ph||20≤2(||p−phk||20+||ph−phk||20). |
Thus obtaining the result
η2Th(Th)−2η2Thk(Thk)≤2Cη2Th(Rh), | (83) |
with Lemma 5.1. By applying the dominance property which is similar to Remark 2.1 in [8], we infer that
osc2Th(T)≤η2Th(T), | (84) |
for
osc2Th(f,T)≤η2Thk(f,T), | (85) |
for
osc2Th(Th∖Rh)−2osc2Thk(Th∖Rh)≤2C(||yh−yhk||20+||ph−phk||20). | (86) |
Then it can be derived via using (83), (84) and (86) that
C(1−2δ)η2Th(Th)≤(1+4C)η2Th(Rh), | (87) |
which tells the proof.
Lemma 5.3. Assume that the marking parameter
#Mhk≤Cβ12sδ−12s|(u,y,p,yd)|1ss(e2Thk+osc2Thk(Thk))−12s, |
if
Proof. Let
#Thν−#Th0≤C|(u,y,p,yd)|1sν−1s, | (88) |
||u−uhν||20+||y−yhν||20+||p−phν||20+osc2Thν(f,Thν)+osc2Thν(yhν−yd,Thν)≤ν2, | (89) |
for any
e2Thϵ+osc2Thϵ(Thϵ)≤β(e2Thν+osc2Thν(Thν), | (90) |
where there are some notations been defined by
e2Thν=||u−uhν||20+||y−yhν||20+||p−phν||20,osc2Thν(Thν)=osc2Thν(f,Thν)+osc2Thν(yhν,Thν), |
and apparently
||v−vhν||20=||v−vhϵ||20+||vhϵ−vhν||20+2(v−vhϵ,vhϵ−vhν), | (91) |
for all
(u−uhϵ,uhϵ−uhν)=(u−uhν,uhϵ−uhν)+(uhν−uhϵ,uhϵ−uhν)≤(u−uhν,uhϵ−uhν)≤||u−uhν||20+14||uhϵ−uhν||20, | (92) |
and apparently in the same way
(y−yhϵ,yhϵ−yhν)=||y−yhν||20+14||yhϵ−yhν||20, | (93) |
(p−phϵ,phϵ−phν)=||p−phν||20+14||phϵ−phν||20. | (94) |
Hence combining (91)-(94) to get
||u−uhϵ||20+||uhϵ−uhν||20+||y−yhϵ||20+||yhϵ−yhν||20+||p−phϵ||20+||phϵ−phν||20≤6(||u−uhν||20+||y−yhν||20+||p−phν||20). | (95) |
For all
∑T∈ThT′||f−f2T||20,T=∑T∈ThT′(∫Tf2−(∫Tf)2|T|)=∫T′f2−∑T∈ThT′(∫Tf)2|T|≤∫T′f2−∑T∈ThT′(∫Tf)2|T′|≤C∫T′f2−(∫T′f)2|T′|=C||f−fT||20,T′, |
which tells that
osc2Thϵ(f,Thϵ)≤Cosc2Thν(f,Thν). | (96) |
By using the same way, we have
osc2Thϵ(yhϵ−yd,Thϵ)≤Cosc2Thν(yhν−yd,Thν). | (97) |
Therefore, we can gain (90) via employing (95)-(97). Based on the definition of
e2Thϵ+osc2Thϵ(Thϵ)≤βν2=δ(e2Thk+osc2Thk(Thk)). |
According to Lemma 5.2, we find that the subset
#Mhk≤#RThk→Thϵ≤#Thϵ−#Thk≤#Thν−Th0. | (98) |
Hence we can obtain the desired result by contacting with (88), (98), and the definition of
Next we derive a equivalent property of the indicator dominates oscillation by concluding Theorem 3.2, Lemma 3.1, and Lemma 4.1 as follows
e2Thk+osc2Thk≈e2hk+γη2Thk(Thk), | (99) |
which is of vital importance for the proof of quasi-optimality.
Theorem 5.4. Assume that
#Thk−#Th0≤C|(u,y,p,yd)|1ss(e2Thk+osc2Thk(Thk))−12s, |
provided
Proof. Similar to the Theorem 4 in [14], we infer that
#Thk−#Th0≤Ck−1∑i=0Mhk. | (100) |
Employing (100) and Lemma 5.3 to get that
#Thk−#Th0≤Ck−1∑i=0Mhk≤Cχk−1∑1=0(e2Thk+osc2Thk(Thk))−12s, | (101) |
where
#Thk−#Th0≤Cχ(e2Thk+osc2Thk(Thk))−12sk∑i=0αik, |
which yields the proof of Theorem 5.4.
In this section, we firstly present the adaptive iteration method where the purpose is to provide empirical analysis for our theory.
Algorithm 6.1. Given an initial control
a(ykh,wh)+(ϕ(ykh),wh)=(f+uk−1h,wh),∀ wh∈Vh,a(qh,pk−1h)+(ϕ(ykh)pk−1h,qh)=(ykh−yd,qh),∀ qh∈Vh,(αukh+pk−1h,vh−ukh)≥0,∀ vkh∈Uhad, |
and apparently
ukh=1α(−Phpkh+max(0,ˉpkh)), |
where
Example 1. We consider the nonlinear optimal control problem subject to the state equation
−Δy+y3=f+u,−Δp+3y2p=y−yd, |
where we choose
u=1α(max(0,ˉp)−p),p=−(sin(πx1)+sin(πx2)),y=sin(πx1)sin(πx2). |
By simple calculation we have
In terms of the same error and actuarial accuracy, the adaptive refinement process saves time than the uniform refinement process. In Figures 1-2, we provide the profiles of the exact state variables, the numerical state variables, the exact co-state and the co-state on adaptively refined grids with
In Figure 4, we plot the profiles of the numerical state and the co-state variables on uniformly refined grids (
In Figure 6, we show the convergence history of the total error estimate indicators, where we plot the adaptive triangle iterations of 15 adaptive loops with the coefficient
Example 2. We consider the same nonlinear optimal control problem as Example 1 with
y={5×1010e1m,m<0,0,m≥0,p={−7×1010e1m,m<0,0,m≥0, |
where
In Figure 7, we plot the numerical state and the co-state on adaptively refined grids with
We give the comparisons of convergence history of Example 2 in Figure 9. The left plot in Figure 9 is adaptively refined with D
The authors express their thanks to the referees for their helpful suggestions, which led to improvements of the presentation.
1. | Liming Guo, Chunjia Bi, Convergence and quasi-optimality of an adaptive finite element method for nonmonotone quasi-linear elliptic problems on L2 errors, 2023, 139, 08981221, 38, 10.1016/j.camwa.2023.03.005 |