Exercises to Wissenschaftliches Rechnen I/Scientific Computing I (V3E1/F4E1)
Winter 2016 / 17
Prof. Dr. Martin Rumpf
Alexander Effland — Stefanie Heyden — Stefan Simon — Sascha Tölkes
Problem sheet 7
Please hand in the solutions on Tuesday December13!
Exercise21 6Points
LetΩ
⊂
Rdbe a polygonal domain with d≤
3 andT
h be a regular triangulation of Ω. Further, letVh⊂
H01,2(
Ω)
be a finite element space with P1⊂
P, s.t. all elementsˆ are affine equivalent to a reference element, andI
h be the Lagrange interpolation operator. Consider the bilinear forma : H01,2×
H01,2→
R, a(
w,v) =
RΩ
∇
w· ∇
vdx.For f, fh
∈
L2(
Ω)
let u∈
H01,2(
Ω)
be the solution of a(
u,v) =
RΩ f
·
vdx for all v∈
H01,2(
Ω)
, and uh∈
Vh be the discrete solution of a(
uh,vh) =
RΩ fh
·
vhdx for all vh∈
Vh.Prove the a posteriori error estimate
k
u−
uhk
0,2,Ω≤
Ck
f−
fhk
0,2,Ω+
C∑
T∈Th
µ2T
!12 ,
where
µ2T :
= k
h(
T)
2(
fh+
∆uh)k
20,2,T+ ∑
E∈E0(T)
k
h(
T)
32[∇
uh·
n]
Ek
20,2,E.Hint: Consider for w
=
u−
uh a dual solution ϕw s.t.a
(
v,ϕw) =
RΩwvdx for all v
∈
H01,2(
Ω)
.Make use of the regularity estimate
k
ϕwk
2,2,Ω≤
ck
wk
0,2,Ω and the interpolation esti- matesk
ϕw− I
hϕwk
0,2,T≤
ch(
T)
2k
ϕwk
2,2,T andk
ϕw− I
hϕwk
0,2,E≤
ch(
T)
32k
ϕwk
2,2,T (see Exercise 19).Exercise22 4+2+4=10Points Let I
= (
0, 1)
be the unit interval. Consider the bilinear form a(
u,v) =
RIu0v0dx.
(i) Let for x
∈
I the Dirac distribution δx defined byh
δx,φi
:=
RIφ
(
y)
dδx(
y)
:=
φ(
x)
for all φ∈
Cc∞(
I)
.Find a function Gx
∈
H01(
I)
s.t.a
(
v,Gx) =
ZIv
(
y)
dδx(
y)
for all v∈
H01(
I)
.Hint: Consider the intervals
(
0,x)
and(
x, 1)
and make the ansatz thatGx is piece-wise smooth.(ii) Let
T
h be an equidistant mesh on I with grid size h. Furthermore, letVh be the space of piece-wise affine and continuous finite elements w.r.t.T
h, and letI
hbe the corresponding Lagrange interpolation operator. Prove the following estimate for the function Gx:k
Gx− I
hGxk
0,1,I≤
ch2.Attention: Take into account that in generalx is not a multiple ofh.
(iii) For f, fh
∈
L2(
I)
let u∈
H01,2(
I)
be the solution of a(
u,v) =
RI f
·
vdx for all v∈
H1,20(
I)
, and uh∈
Vh be the discrete solution of a(
uh,vh) =
RI fh
·
vhdx for all vh∈
Vh.Prove the a posteriori error estimate
k
u−
uhk
0,∞,I≤
Ck
f−
fhk
0,∞,I+
Ch2∑
T∈Th
k(
fh+
∆uh)k
0,∞,T! .
Hint: Use that a