On Evolution Equations Having Tahar Haddad and Chems - - PowerPoint PPT Presentation

on evolution equations having
SMART_READER_LITE
LIVE PREVIEW

On Evolution Equations Having Tahar Haddad and Chems - - PowerPoint PPT Presentation

On Evolution Equations Having Hypo- monotonicities of Opposite Sign On Evolution Equations Having Tahar Haddad and Chems Hypomonotonicities of Opposite Sign Eddine Arroud Introduction Preliminaries Tahar Haddad and Chems Eddine Arroud


slide-1
SLIDE 1

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

On Evolution Equations Having Hypomonotonicities of Opposite Sign

Tahar Haddad and Chems Eddine Arroud

Departement of Mathematics, University of Jijel, Algeria Padova University, Italy

september 25-29,2017

slide-2
SLIDE 2

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Outline

1

Introduction

slide-3
SLIDE 3

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Outline

1

Introduction

2

Preliminaries

slide-4
SLIDE 4

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Outline

1

Introduction

2

Preliminaries

3

Main result

slide-5
SLIDE 5

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Outline

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

slide-6
SLIDE 6

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Outline

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

5

Bibliographie

slide-7
SLIDE 7

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

Sweeping processes were proposed and thoroughly studied by J.J. Moreau in the seventies,

slide-8
SLIDE 8

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

Sweeping processes were proposed and thoroughly studied by J.J. Moreau in the seventies, As a partial viewpoint, consider a time-moving closed convex set C(t) which drags a point u(t), so this point must stay in C(t) at every time t, and the

  • pposite of its velocity, say −du

dt (t), has to be normal to the set

C(t).

slide-9
SLIDE 9

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

Sweeping processes were proposed and thoroughly studied by J.J. Moreau in the seventies, As a partial viewpoint, consider a time-moving closed convex set C(t) which drags a point u(t), so this point must stay in C(t) at every time t, and the

  • pposite of its velocity, say −du

dt (t), has to be normal to the set

C(t). To take into account the nonsmoothness of the boundary

  • f the convex set C(t), the law of motion is formulated as

  

du dt (t) ∈ −NC(t)(u(t)), a.e.t ∈ [0, T]

u(0) = u0 ∈ C(0) u(t) ∈ C(t) ∀t ∈ [0, T], (1) where NC(t)(·) denotes the normal cone operator of the convex set C(t) in the sense of convex analysis in a Hilbert space H.

slide-10
SLIDE 10

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Existence and uniqueness of solutions of such systems and their classical variants : ⊲ subjected to perturbation forces, ⊲ non convex prox-regular sets C, ⊲ state dependent set C(t, x) ⊲ second-order sweeping processes ... have been considered by many authors in the literature.

slide-11
SLIDE 11

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

On the other hand, in [A. Bressan, A. Cellina, and G. Colombo ;’89] ˙ x(t) ∈ F(x(t)), F(x) ⊂ ∂g(x), x(0) = x0, (2) F is a monotonic upper semicontinuous ( not necessarily convex valued, hence not maximal) map contained in the subdifferential of a locally bounded convex function.

slide-12
SLIDE 12

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

On the other hand, in [A. Bressan, A. Cellina, and G. Colombo ;’89] ˙ x(t) ∈ F(x(t)), F(x) ⊂ ∂g(x), x(0) = x0, (2) F is a monotonic upper semicontinuous ( not necessarily convex valued, hence not maximal) map contained in the subdifferential of a locally bounded convex function.

  • The lack of the minus sign in (2) yields that the distance

among solutions increases and typically there is no uniqueness.

slide-13
SLIDE 13

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Introduction

On the other hand, in [A. Bressan, A. Cellina, and G. Colombo ;’89] ˙ x(t) ∈ F(x(t)), F(x) ⊂ ∂g(x), x(0) = x0, (2) F is a monotonic upper semicontinuous ( not necessarily convex valued, hence not maximal) map contained in the subdifferential of a locally bounded convex function.

  • The lack of the minus sign in (2) yields that the distance

among solutions increases and typically there is no uniqueness.

  • Existence of solutions depends on arguments of convex
  • analysis. This result has been generalized by many authors in

different ways.

slide-14
SLIDE 14

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

The dynamics

The purpose of the present talk is to study a Cauchy problem ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, (3) ⊲ in a infinite dimensional Hilbert space H,

slide-15
SLIDE 15

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

The dynamics

The purpose of the present talk is to study a Cauchy problem ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, (3) ⊲ in a infinite dimensional Hilbert space H, ⊲ the closed set C is locally prox-regular at x0 ( hence NC is hypomonotone set valued mapping),

slide-16
SLIDE 16

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

The dynamics

The purpose of the present talk is to study a Cauchy problem ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, (3) ⊲ in a infinite dimensional Hilbert space H, ⊲ the closed set C is locally prox-regular at x0 ( hence NC is hypomonotone set valued mapping), ⊲ the set valued mapping F is not necessarily the whole subdifferential of g , and we take the plus sign, instead of the classical minus.

slide-17
SLIDE 17

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

The dynamics

The purpose of the present talk is to study a Cauchy problem ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, (3) ⊲ in a infinite dimensional Hilbert space H, ⊲ the closed set C is locally prox-regular at x0 ( hence NC is hypomonotone set valued mapping), ⊲ the set valued mapping F is not necessarily the whole subdifferential of g , and we take the plus sign, instead of the classical minus. ⊲ The system (3) can be considered as a non convex cyclical monotone differential inclusion under control term u(t) ∈ −NC(·) which guarantees that the trajectory x(t) always belongs to the desired set C for all t ∈ [0, T]. We prove (local) existence of solutions.

slide-18
SLIDE 18

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

In [Castaing, C., Syam, ’03], the authors proved the existence of solutions of problem (3), ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C,

slide-19
SLIDE 19

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

In [Castaing, C., Syam, ’03], the authors proved the existence of solutions of problem (3), ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C,

  • the closed set C was supposed to be compact and

uniformly r-prox-regular,

slide-20
SLIDE 20

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

It is also worth mentioning that problem (3) in the more general form ˙ x(t) ∈ F(x(t))−∂V (x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ domV , (4)

  • g is ϕ-convex of order two
  • V has aϕ-monotone subdifferential of order two (shortly

V ∈ MS(2)) has been studied by : [Cardinali, T., Colombo, G., Papalini, F., Tosques, M., Nonlinear’97]

slide-21
SLIDE 21

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

It is also worth mentioning that problem (3) in the more general form ˙ x(t) ∈ F(x(t))−∂V (x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ domV , (4)

  • g is ϕ-convex of order two
  • V has aϕ-monotone subdifferential of order two (shortly

V ∈ MS(2)) has been studied by : [Cardinali, T., Colombo, G., Papalini, F., Tosques, M., Nonlinear’97] Notice that we can obtain (3) from (4), by taking V = δC, the indicator function of the set C, i.e. δC(x) = 0 for X ∈ C and δC(x) = ∞ for x ∈ H \ C.

slide-22
SLIDE 22

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

It is also worth mentioning that problem (3) in the more general form ˙ x(t) ∈ F(x(t))−∂V (x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ domV , (4)

  • g is ϕ-convex of order two
  • V has aϕ-monotone subdifferential of order two (shortly

V ∈ MS(2)) has been studied by : [Cardinali, T., Colombo, G., Papalini, F., Tosques, M., Nonlinear’97] Notice that we can obtain (3) from (4), by taking V = δC, the indicator function of the set C, i.e. δC(x) = 0 for X ∈ C and δC(x) = ∞ for x ∈ H \ C. Indeed, as C is locally pox-reqular, then V = δC is pln ( proposition 3.3 in [M.Mazade ad L.Thibault, ’12) and so MS(2).

slide-23
SLIDE 23

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Despite the similarity of (3) and (4), the problems are quite different, since in general with V = δC the level set {x ∈ H; V (x) ≤ r} is not compact, and these were basic assumption in previous works.

slide-24
SLIDE 24

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Despite the similarity of (3) and (4), the problems are quite different, since in general with V = δC the level set {x ∈ H; V (x) ≤ r} is not compact, and these were basic assumption in previous works. Since these condition do not holds for (3), she has to be replaced by suitable substitute in case of sweeping processes.

slide-25
SLIDE 25

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Despite the similarity of (3) and (4), the problems are quite different, since in general with V = δC the level set {x ∈ H; V (x) ≤ r} is not compact, and these were basic assumption in previous works. Since these condition do not holds for (3), she has to be replaced by suitable substitute in case of sweeping processes. In this talk, we give a new approach, in which the compactness assumption is shifted from the set C to the set-valued map F.

slide-26
SLIDE 26

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

5

Bibliographie

slide-27
SLIDE 27

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

x − → NC(x)... weak smoothness

  • While one would like to consider very general sets C there are

limits to possible sets on which the existence of solutions (well-posedness ) of sweeping processes ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, can be developed. ‘

slide-28
SLIDE 28

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

x − → NC(x)... weak smoothness

  • While one would like to consider very general sets C there are

limits to possible sets on which the existence of solutions (well-posedness ) of sweeping processes ˙ x(t) ∈ F(x(t))−NC(x(t)), F(x) ⊂ ∂g(x), x(0) = x0 ∈ C, can be developed. ‘

  • Namely, For a fixed closed subset C, the set-valued map

x − → NC(x) is not upper semicontunous which is needed for the proof of existence of solutions.

slide-29
SLIDE 29

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Uniformly r− prox-regular set

  • We say that C is uniformly r-prox-regular provided the

inequality v, y − x ≤ 1 2ry − x2 holds ∀x, y ∈ C, ∀v the unit external normal to C.

slide-30
SLIDE 30

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Uniformly r− prox-regular set

  • We say that C is uniformly r-prox-regular provided the

inequality v, y − x ≤ 1 2ry − x2 holds ∀x, y ∈ C, ∀v the unit external normal to C.

  • an external tangent ball with radius smaller than 1

2r

depending on the tangency point x, can be rolled around C with it’s closure touching C only at x.

  • In particular uniformly r−prox-regular sets can have outside

corners and outside cusps,but not inside corners.

slide-31
SLIDE 31

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

locally prox-regular set

Definition For positive real numbers r and α, the closed set C is said to be(r, α) prox-regular at a point x ∈ C provided that for any x ∈ C ∩ B(x, α) and any v ∈ NP

C (x) such that v ≤ r, one

has x = projC(x + v). The set C is said to be r-uniformly prox-regular when α = +∞.

slide-32
SLIDE 32

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

primal lower nice function

Let us now define primal lower nice function in a quantified way [Mazade and Thibault ’12]. Definition Let f : H → R ∪ {+∞} be a proper lower semicontinuous

  • function. The function f is said to be primal lower nice (pln, for

short) on an open convex set O with O ∩ domf = ∅ if there exists some real number c ≥ 0 such that for all x ∈ O ∩ dom∂pf (x) and for all v ∈ ∂pf (x) on has f (y) ≥ f (x) + v, y − x − c(1 + v)y − x2, (5) for each y ∈ O. The real c ≥ 0 will be called a pln constant for f over O and we will say that f is c -pln on O.

slide-33
SLIDE 33

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Auxiliary existence result

We have the following existence and uniqueness result [Mazade and Thibault ’12]. Let C be an (r, α)-prox-regular set at the point x0 ∈ C and let any real number η0 ∈]0, α[. Then for any x ∈ B(x0, α − η0) ∩ C, any positive real number τ ≤ T0 − T, and any mapping h ∈ L1([0, T], H) such that T0+τ

T0

h(s)ds < η0/2, the differential variational inequality ˙ x(t) ∈ −NC(x(t)) + h(t), x(0) = x, a.e t ∈ [T0, T0 + τ], (6) has an absolutely continuous solution x : [T0, T0 + τ] → B(x, η0) ∩ C. Moreover, ˙ x(t) ≤ ˙ x(t) − h(t) + h(t) ≤ 2h(t) a.e t ∈ [T0, T0 + τ].

slide-34
SLIDE 34

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

necessarily condition of existence

With the previous assumptions we proved the following property of solution. Lemma The absolutely continuous solution x : [T0, T0 + τ] → B(x, η0) ∩ C of ˙ x(t) ∈ −NC(x(t)) + h(t), x(0) = x, a.e t ∈ [T0, T0 + τ], satisfies the following property ˙ x(t), ˙ x(t) = h(t), ˙ x(t) a.e t ∈ [T0, T0 + τ].

slide-35
SLIDE 35

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

5

Bibliographie

slide-36
SLIDE 36

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Main result

we prove the existence of local solutions to the evolution problem ˙ x(t) ∈ F(x(t))−NC(x(t)), x(0) = x0 ∈ C, a.e t ∈ [0, T], (7) Under the following assumptions :

slide-37
SLIDE 37

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Main result

we prove the existence of local solutions to the evolution problem ˙ x(t) ∈ F(x(t))−NC(x(t)), x(0) = x0 ∈ C, a.e t ∈ [0, T], (7) Under the following assumptions : (HC1) : the closed set C is (r, α)-prox-regular at the point x0 ∈ C ;

slide-38
SLIDE 38

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Main result

we prove the existence of local solutions to the evolution problem ˙ x(t) ∈ F(x(t))−NC(x(t)), x(0) = x0 ∈ C, a.e t ∈ [0, T], (7) Under the following assumptions : (HC1) : the closed set C is (r, α)-prox-regular at the point x0 ∈ C ; (HF1) : O ⊂ H is an open convex set containing B(x0, η0) and F : O → 2H is an upper semicontinous set-valued mapping with nonempty weakly compact values ;

slide-39
SLIDE 39

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Main result

we prove the existence of local solutions to the evolution problem ˙ x(t) ∈ F(x(t))−NC(x(t)), x(0) = x0 ∈ C, a.e t ∈ [0, T], (7) Under the following assumptions : (HC1) : the closed set C is (r, α)-prox-regular at the point x0 ∈ C ; (HF1) : O ⊂ H is an open convex set containing B(x0, η0) and F : O → 2H is an upper semicontinous set-valued mapping with nonempty weakly compact values ; (HF2) : let g : O → R ∪ {+∞} be a proper lower semicontinous function c-pln on O with F(x) ⊂ ∂Cg(x), ∀x ∈ O such that g is locally bounded from above on O.

slide-40
SLIDE 40

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

By a solution of inclusion (7) we mean an absolutely continuous function x(.) : [0, T] → H, x(0) = x0 ∈ C, such that the inclusion ˙ x(t) ∈ −NC(x(t))+f (t) holds a.e. for some f ∈ L2([0, T], H) such that f (t) ∈ F(x(t)) a.e.

slide-41
SLIDE 41

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

By a solution of inclusion (7) we mean an absolutely continuous function x(.) : [0, T] → H, x(0) = x0 ∈ C, such that the inclusion ˙ x(t) ∈ −NC(x(t))+f (t) holds a.e. for some f ∈ L2([0, T], H) such that f (t) ∈ F(x(t)) a.e. Theorem Assume that H is the Hilbert space, (HC1), (HF1) and (HF2) : F(x) ⊂ K for all x ∈ C with K strongly compact in H hold. Then, there exists T > 0 and an absolutely continuous function x(.) : [0, T] B(x0, η0) a local solution to problem (7).

slide-42
SLIDE 42

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Sketch of the proof

⊲ Let x0 ∈ C and and let g : O → R ∪ {+∞} satisfy (HF2).

slide-43
SLIDE 43

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Sketch of the proof

⊲ Let x0 ∈ C and and let g : O → R ∪ {+∞} satisfy (HF2). ⊲ F(x) ⊂ ∂pg(x) it follows that F is bounded by M on B(x0, η0).

slide-44
SLIDE 44

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Sketch of the proof

⊲ Let x0 ∈ C and and let g : O → R ∪ {+∞} satisfy (HF2). ⊲ F(x) ⊂ ∂pg(x) it follows that F is bounded by M on B(x0, η0). ⊲ Let T > 0 such that T < min{η0/2M, (α − η0)/2M}, where α and η0 are given by (HC1).

slide-45
SLIDE 45

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Sketch of the proof

⊲ Let x0 ∈ C and and let g : O → R ∪ {+∞} satisfy (HF2). ⊲ F(x) ⊂ ∂pg(x) it follows that F is bounded by M on B(x0, η0). ⊲ Let T > 0 such that T < min{η0/2M, (α − η0)/2M}, where α and η0 are given by (HC1). Our purpose is to prove that there exists x : [0, T] → B(x0, η0) ∩ C a solution to the Cauchy problem (7)

slide-46
SLIDE 46

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Construction of approximates solutions

xn

k (.) : [tn k , tn k+1] B(x0, η0) ∩ C with 0 ≤ k ≤ n − 1, such

that for each k ∈ 0, ..., n − 1 (with xn

−1(tn 0) = x0)

˙ xn

k (t) ∈ −NC(xn k (t)) + f n k (t),

xn

k (tn k ) = xn k−1(tn k ), t ∈ [tn k , tn k+1],

f n

k (t) ∈ F(xn k−1(tn k ))

and ˙ (xn

k )(t) ≤ 2M, a.e

t ∈ [tn

k , tn k+1]

slide-47
SLIDE 47

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

sequence of approximates solutions

Define xn(.), fn(.) : [0, ¯ T] → H by xn(t) =

n−1

  • k=0

xn

k (t)χ[tn

k ,tn k+1](t),

fn(t) =

n−1

  • k=0

f n

k (t)χ(tn

k ,tn k+1](t)

where χA is the characteristic function of the set A. The mapping xn(.) is absolutely continuous on [0, ¯ T] with xn(t) ∈ B(x0, η0) ∩ C; ∀t ∈ [0, ¯ T] (8) Further, putting θn(t) := tn

k

if t ∈ [tn

k , tn k+1[

for k ∈ 0, ..., n − 1, θn( ¯ T) = (9) we have ˙ xn(t) ∈ −NC(xn(t))+fn(t), xn(0) = x0, a.e t ∈ [0, ¯ T] (10)

slide-48
SLIDE 48

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence of sequences

Let us define Zn(t) := t

0 fn(s)ds.

Then for all t ∈ [0, T] the set {Zn(t), n ∈ N∗} is contained in the strong compact set T co({0} ∪ K) and so it is relatively strongly compact in H.

slide-49
SLIDE 49

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence of sequences

Let us define Zn(t) := t

0 fn(s)ds.

Then for all t ∈ [0, T] the set {Zn(t), n ∈ N∗} is contained in the strong compact set T co({0} ∪ K) and so it is relatively strongly compact in H. ⊲ Then by Arzela Asccoli’s theorem we get the relative compactness of the set {Zn, n ∈ N∗} with respect to the uniform convergence in C([0, T], H) and so we may assume that without loss of generality that (Zn)n converges uniformly to some mapping Z.

slide-50
SLIDE 50

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence of sequences

Let us define Zn(t) := t

0 fn(s)ds.

Then for all t ∈ [0, T] the set {Zn(t), n ∈ N∗} is contained in the strong compact set T co({0} ∪ K) and so it is relatively strongly compact in H. ⊲ Then by Arzela Asccoli’s theorem we get the relative compactness of the set {Zn, n ∈ N∗} with respect to the uniform convergence in C([0, T], H) and so we may assume that without loss of generality that (Zn)n converges uniformly to some mapping Z. ⊲ As fn(t) ≤ M, we may suppose that (fn)n converges weakly in L2([0, T], H) to some mapping f .

slide-51
SLIDE 51

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence of sequences

Let us define Zn(t) := t

0 fn(s)ds.

Then for all t ∈ [0, T] the set {Zn(t), n ∈ N∗} is contained in the strong compact set T co({0} ∪ K) and so it is relatively strongly compact in H. ⊲ Then by Arzela Asccoli’s theorem we get the relative compactness of the set {Zn, n ∈ N∗} with respect to the uniform convergence in C([0, T], H) and so we may assume that without loss of generality that (Zn)n converges uniformly to some mapping Z. ⊲ As fn(t) ≤ M, we may suppose that (fn)n converges weakly in L2([0, T], H) to some mapping f . ⊲ Then for all t ∈ [0, T] Z(t) = limnZn(t) = limn t fn(s)ds = t f (t)dt.

slide-52
SLIDE 52

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence of sequences

Let us define Zn(t) := t

0 fn(s)ds.

Then for all t ∈ [0, T] the set {Zn(t), n ∈ N∗} is contained in the strong compact set T co({0} ∪ K) and so it is relatively strongly compact in H. ⊲ Then by Arzela Asccoli’s theorem we get the relative compactness of the set {Zn, n ∈ N∗} with respect to the uniform convergence in C([0, T], H) and so we may assume that without loss of generality that (Zn)n converges uniformly to some mapping Z. ⊲ As fn(t) ≤ M, we may suppose that (fn)n converges weakly in L2([0, T], H) to some mapping f . ⊲ Then for all t ∈ [0, T] Z(t) = limnZn(t) = limn t fn(s)ds = t f (t)dt. ⊲ then Z is absolutely continuous and ˙ Z(t) = f (t) for almost t ∈ [0, T].

slide-53
SLIDE 53

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

(xn)n is a Cauchy sequence in the space C([0, T], H)

Fix m, n ∈ N∗. Put εm,n := Zm − Zn∞ → 0, and wn(t) := xn(t) − Zn(t), we

  • btain
slide-54
SLIDE 54

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

(xn)n is a Cauchy sequence in the space C([0, T], H)

Fix m, n ∈ N∗. Put εm,n := Zm − Zn∞ → 0, and wn(t) := xn(t) − Zn(t), we

  • btain

d dt (wm(t) − wn(t)2) ≤ 2M r wn(t) − wm(t)2 + γεm,n, where γ is some positive constant independent of m, n and t.

slide-55
SLIDE 55

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

(xn)n is a Cauchy sequence in the space C([0, T], H)

Fix m, n ∈ N∗. Put εm,n := Zm − Zn∞ → 0, and wn(t) := xn(t) − Zn(t), we

  • btain

d dt (wm(t) − wn(t)2) ≤ 2M r wn(t) − wm(t)2 + γεm,n, where γ is some positive constant independent of m, n and t. As wm(0) − wn(0) = 0, the Gronwall inequality implies for almost t wm(t) − wn(t)2 ≤ Lεm,n, L is some positive constant independent of m, n and t. Hence (wn)n converges uniformly to some mapping w and so (xn)n converges uniformly to some mapping x := w + Z.

slide-56
SLIDE 56

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

We prove that ˙ x(t) ∈ −NC(x(t)) + f (t), a.e t ∈ [0, ¯ T]

By construction we have −˙ xn(t) + fn(t) ∈ NC(xn(t)) ∩ B(0, M) = M∂dC(xn(t)). Since (−˙ xn + fn)n converges weakly in L2([0, ¯ T], H) and (xn)n converges strongly in C([0, ¯ T], H) By invoking a closure type Lemma

slide-57
SLIDE 57

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

We prove that ˙ x(t) ∈ −NC(x(t)) + f (t), a.e t ∈ [0, ¯ T]

By construction we have −˙ xn(t) + fn(t) ∈ NC(xn(t)) ∩ B(0, M) = M∂dC(xn(t)). Since (−˙ xn + fn)n converges weakly in L2([0, ¯ T], H) and (xn)n converges strongly in C([0, ¯ T], H) By invoking a closure type Lemma −˙ x(t) + f (t) ∈ M∂CdC(x(t)), a.e t ∈ [0, ¯ T]. (11) The last inclusion and x(t) ∈ C ensure    ˙ x(t) ∈ −NC(x(t)) + f (t), a.e t ∈ [0, ¯ T] x(t) ∈ B(x0, η0) ∩ C; ∀t ∈ [0, ¯ T] x(0) = x0 ∈ C, (12)

slide-58
SLIDE 58

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Convergence for the norm topology of L2

H of the

sequence (˙ xn(·))

  • By exploiting that g is c − pln,
  • Integrating

˙ x(t), ˙ x(t) = f (t), ˙ x(t). and ˙ xn(t), ˙ xn(t) = fn(t), ˙ xn(t). we obtain lim sup

n

T ˙ xn(t)2dt ≤ T ˙ x(t)2dt. By the weak lower semicontinuity of the norm, we deduce that lim

n

T ˙ xn(t)2dt = T ˙ x(t)2dt, which implies that (˙ xn)n converges to ˙ x in the strong topology L2([0, ¯ T], H). Therefore, there exists a subsequence, still denoted by (˙ xn)n which converges pointwise a.e. to ˙ x.

slide-59
SLIDE 59

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

˙ x(t) ∈ −NC(x(t)) + F(x(t)), a.e t ∈ [0, ¯ T].

Define the set-valued mapping G(x) := −M∂dC(x), x ∈ H. Fix t ∈ [0, ¯ T] \ N, we have (d∗(A, B) := sup{d(a, B) : a ∈ A} for A, B ⊂ H)

slide-60
SLIDE 60

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

˙ x(t) ∈ −NC(x(t)) + F(x(t)), a.e t ∈ [0, ¯ T].

Define the set-valued mapping G(x) := −M∂dC(x), x ∈ H. Fix t ∈ [0, ¯ T] \ N, we have (d∗(A, B) := sup{d(a, B) : a ∈ A} for A, B ⊂ H) d(˙ xn(t), G(x(t))+F(x(t)) ≤ d∗(G(xn(t))+F(xn(θn(t))), G(x(t))+ ≤ d∗(G(xn(t)), G(x(t))) + d∗(F(xn(θn(t))), F(x(t))). Since ˙ xn(t) → ˙ x(t), xn(t) → x(t) and ˙ xn(θn(t)) → ˙ x(t) strongly in H, the upper semicontinuity of F and G give d(˙ x(t), G(x(t)) + F(x(t)) = 0.

slide-61
SLIDE 61

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

As x(t) ∈ C we finally get ˙ x(t) ∈ G(x(t)) + F(x(t)) = −M∂pdC(x(t)) + F(x(t)) ⊂ −NC(x(t)) + F(x(t)). This completes the proof of Theorem.

slide-62
SLIDE 62

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

5

Bibliographie

slide-63
SLIDE 63

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Comments and open problem

⊲ In this talk, the compactness assumption is shifted from the set C to the perturbation F. More precisely, we cannot use Arzela-Ascolis theorem . However we prove that (xn) is a Cauchy sequence in the space C([0, T], H). These arguments, used with convex-valued mapping F and uniformly prox-regular set C (see [Bounkhel and Thibault ’05]), still hold for (3). ⊲ The property ˙ x(t), ˙ x(t) = f (t), ˙ x(t) a.e t guarantees the convergence for the norm topology of L2

H of the

sequence (˙ xn(·)).

slide-64
SLIDE 64

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Comments

⊲ We can extent the existence result to the non constant set-valued map C(t) := C + a(t) where C is a fixed closed locally prox-regular subset x0, and a(t) an absolutely continuous mapping from [0, T] into H, with a(0) = 0. Also the previous property becomes h(t) − ˙ x(t), ˙ x(t) − ˙ a(t) = 0 a.e t

slide-65
SLIDE 65

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

  • pen problem

Further, it would be interesting to address the validity of the preceding result to the general time-dependent when C(t) ˙ x(t) ∈ F(x(t)) − NC(t)(x(t)), F(x)⊂ ∂g(x), x(0) = x0 ∈ C(0),

slide-66
SLIDE 66

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

1

Introduction

2

Preliminaries

3

Main result

4

Comments and open problem

5

Bibliographie

slide-67
SLIDE 67

On Evolution Equations Having Hypo- monotonicities

  • f Opposite

Sign Tahar Haddad and Chems Eddine Arroud Introduction Preliminaries Main result Comments and open problem Bibliographie

Bibliography

  • S. Adly, T. Haddad, L. Thibault, Convex sweeping process

in the framework of measure differential inclusions and evolution variational inequalities, Math. Program. Ser. B 148 (2014), 5-47.

  • A. Bressan ; A. Cellina and G. Collombo Upper

semicontuous differential inclusions without convexity, Proc.Am. Math. Soc 106 (1989) 771-775. Bounkhel, M., Thibault, L. : Nonconvex sweeping process and prox regularity in Hilbert space. J. Nonlinear Convex

  • Anal. 6, 359374 (2005)

Cardinali, T., Colombo, G., Papalini, F., Tosques, M. : On a class of evolution equations without convexity, Nonlinear Analysis : Theory, Meth. Appl. 28, 217-234 (1997)

  • M. Kunze and M. D. P. Monteiro Marques, An

introduction to Moreaus sweeping process. In : Brogliato,

  • B. (ed.) Impacts in Mechanical Systems. Analysis and

Modelling, pp. 160. Springer, Berlin (2000) Mazade, M., Thibault, L. : Differential variational inequalities with locally prox regular sets. J. Convex Anal. 19(4), 11091139 (2012) Syam, A., Castaing, C. : On a class of evolution inclusions governed by a nonconvex sweeping process. Nonlinear Analysis and Applications, Vol 1, 2, 341-359, Kluwer Acad.

  • PublI. Dordrecht