Abstract

We prove the existence of solutions for third-order nonconvex state-dependent sweeping process with unbounded perturbations of the form: 𝐴(𝑥(3)(𝑡))𝑁(𝐾(𝑡,̇𝑥(𝑡));𝐴(̈𝑥(𝑡)))+𝐹(𝑡,𝑥(𝑡),̇𝑥(𝑡),̈𝑥(𝑡))+𝐺(𝑥(𝑡),̇𝑥(𝑡),̈𝑥(𝑡))a.e.[0,𝑇], 𝐴(̈𝑥(𝑡))𝐾(𝑡,̇𝑥(𝑡)), a.e.𝑡[0,𝑇], 𝑥(0)=𝑥0,̇𝑥(0)=𝑢0, ̈𝑥(0)=𝜐0, where 𝑇>0, 𝐾 is a nonconvex Lipschitz set-valued mapping, 𝐹 is an unbounded scalarly upper semicontinuous convex set-valued mapping, and 𝐺 is an unbounded uniformly continuous nonconvex set-valued mapping in a separable Hilbert space .

1. Introduction

In the seventies, Moreau introduced and studied in [1] the following differential inclusion:̇𝑥(𝑡)𝑁𝐾(𝑡)𝑥(𝑥(𝑡))a.e.on𝐼,(0)=𝑥0𝐾(0),(SP) where 𝐼=[0,𝑇](𝑇>0),𝐾𝐼 is a set-valued mapping defined from 𝐼 to a Hilbert space and takes closed convex values, 𝑁𝐾(𝑡)(𝑥(𝑡)) denotes the normal cone to the set 𝐾(𝑡) at the position 𝑥(𝑡). The differential inclusion (SP) is called the Moreau’s sweeping process problem. In [27], the authors studied the existence of solutions for various extensions and variants of (SP). In [6], the author studied for the first time the existence of solutions for the following type of second-order differential inclusions𝑥̈𝑥(𝑡)𝑁(𝐾(𝑥(𝑡));̇𝑥(𝑡)),(0)=𝑥0,̇𝑥(𝑡)𝐾(𝑥(𝑡)),(SSP) with convex set-valued mapping 𝐾. The problem (SSP) has been extended in several ways. For instance, in [8], the authors studied a variant of (SSP) with a perturbation,[],𝐴𝑥(𝑡)𝑁(𝐾(𝑡);̇𝑥(𝑡))+̈𝑥(𝑡)+𝐹(𝑡,̇𝑥(𝑡))a.e.on0,𝑇(1.1) when the set-valued mapping 𝐾 is not necessarily convex and 𝐴 is a linear and bounded operator on a separable Hilbert space. In [9], the author studied existence results for the following general problem:𝑥̈𝑥(𝑡)𝑁(𝐾(𝑥(𝑡));̇𝑥(𝑡))+𝐹(𝑡,𝑥(𝑡),̇𝑥(𝑡))+𝐺(𝑡,𝑥(𝑡),̇𝑥(𝑡)),a.e.𝐼,(0)=𝑥𝑜,̇𝑥(0)=𝑢0,̇𝑥(𝑡)𝐾(𝑥(𝑡)),𝑡𝐼,(SSPMP) where 𝐾 is a nonconvex set-valued mapping, 𝐹 is an unbounded scalarly upper semicontinuous convex set-valued mapping, and 𝐺 is an unbounded continuous nonconvex set-valued mapping in a separable Hilbert space . The differential inclusion (SSPMP) is called Second-order Sweeping Process with Mixed Perturbations. In this work, we prove the existence of solutions for the following form of third-order differential inclusions (TSPMP):𝐴𝑥(3)(𝑡)𝑁(𝐾(𝑡,̇𝑥(𝑡));𝐴(̈𝑥(𝑡)))+𝐹(𝑡,𝑥(𝑡),̇𝑥(𝑡),𝐴(̈𝑥(𝑡)))+𝐺(𝑡,𝑥(𝑡),̇𝑥(𝑡),𝐴(̈𝑥(𝑡)))a.e.on𝐼,𝐴(̈𝑥(𝑡))𝐾(𝑡,̇𝑥(𝑡)),a.e.𝑡𝐼,𝑥(0)=𝑥0,̇𝑥(0)=𝑢0,̈𝑥(0)=𝜐0,(TSPMP) where 𝐾 is a nonconvex Lipschitz set-valued mapping, 𝐴 is surjective bounded linear operator, 𝐹 is an unbounded scalarly upper semicontinuous convex set-valued mapping, and 𝐺 is an unbounded uniformly continuous nonconvex set-valued mapping in a separable Hilbert space . We will call it third-order nonconvex state-dependent sweeping process with mixed perturbations (in short (TSPMP)). Problem (TSPMP) includes as a special case the following differential variational inequality (DVI): given a convex compact set 𝐷 in , three points 𝑥0,𝑢0,𝑣0𝐷+𝑓(0,𝑢0) with Λ(𝑣0𝑓(0,𝑢0))0:[]Find𝑇>0andaLipschitzmapping𝑥0,𝑇suchthat(1)𝑥(0)=𝑥0,̇𝑥(0)=𝑢0,̈𝑥(0)=𝑣0;[];𝑥(2)Λ(̈𝑥(𝑡)𝑓(𝑡,̇𝑥(𝑡)))0,̈𝑥(𝑡)𝐷+𝑓(𝑡,̇𝑥(𝑡)),a.e.on0,𝑇(3)𝑤𝐷+𝑓(𝑡,̇𝑥(𝑡))withΛ(𝑤𝑓(𝑡,̇𝑥(𝑡)))0,wehave(3)(𝑡)𝛼𝑥(𝑡),𝑤̈𝑥(𝑡)𝑎(𝑥(𝑡)+̇𝑥(𝑡)̈𝑥(𝑡),𝑤̈𝑥(𝑡))+𝜌𝑤̈𝑥(𝑡)2[],,a.e.on0,𝑇(1.2) where 𝛼>0,Λ is inf-compact and lower-𝐶2 function, 𝑎(,) is a real bilinear, symmetric, bounded, and elliptic form on ×, and 𝑓𝐼× is a Lipschitz function. We use our main theorem to prove that (DVI) has at least one Lipschitz solution.

This paper is organized as follows. Section 2 contains some definitions, notations, and important results needed in the paper. In Section 3, we prove an existence result for (TSPMP) when the set-valued mapping 𝐾 is not necessarily convex, by using ideas and techniques from Nonsmooth Analysis. The result is proved by showing that a sequence of approximate solutions converges to a solution of (TSPMP). Then, we deduce from our main theorem an existence result for a second-order nonconvex differential inclusion ̈𝑥(𝑡)𝐾(𝑡,̇𝑥(𝑡))foralmostall𝑡𝐼, where the right-hand side is not convex. In Section 4, we study the closedness and the compactness of the solution sets of (TSPMP). In Section 5, we state an application to differential variational inequalities (DVI).

2. Preliminaries

Throughout the paper, will denote a Hilbert space. We need to recall some notation and definitions that will be used in all the paper. Let 𝑆 be a nonempty closed subset of . We denote by 𝑑𝑆() the usual distance function to the subset 𝑆, that is, 𝑑𝑆(𝑥)=inf𝑧𝑆𝑥𝑧. We recall (see [10]) that the proximal normal cone of 𝑆 at 𝑥 is given by𝑁𝑃(𝑆;𝑥)={𝜉𝛼>0s.t.𝑥Proj(𝑥+𝛼𝜉;𝑆)},(2.1) where𝑥Proj(𝑥;𝑆)=𝑆𝑑𝑆(𝑥)=𝑥𝑥.(2.2) Equivalently 𝑁𝑃(𝑆;𝑥) can be defined by (2) as the set of all 𝜉 for which there exist 𝜎>0 such that𝜉,𝑥𝑥𝑥𝜎𝑥2,𝑥𝑆.(2.3) Now, let 𝑓{+} be a function and 𝑥 any point in where 𝑓 is finite. We recall that the proximal subdifferential 𝜕𝑃𝑓(𝑥) is the set of all 𝜉 for which there exist 𝛿,𝜎>0 such that for all 𝑥𝑥+𝛿𝔹𝜉,𝑥𝑥𝑥𝑓𝑥𝑓(𝑥)+𝜎𝑥.(2.4) Here 𝔹 denotes the closed unit ball centered at the origin of . Recall that for a given 𝑟]0,+], a subset 𝑆 is uniformly prox-regular with respect to 𝑟 (we will say uniformly 𝑟-prox-regular) (see [10, 11]) if and only if for all 𝑥𝑆 and all 0𝜉𝑁𝑃(𝑆;𝑥) one has𝜉𝜉,𝑥𝑥12𝑟𝑥𝑥2,(2.5) for all 𝑥𝑆. We make the convention 1/𝑟=0 for 𝑟=+. Recall that for 𝑟=+, the uniform 𝑟-prox-regularity of 𝑆 is equivalent to the convexity of 𝑆, which makes this class of great importance.

In [12], the authors established the following characterization of the uniform prox-regularity in terms of the subdifferential of the distance function. We recall here a consequence of their result needed in the sequel.

Proposition 2.1. Let 𝑆 be nonempty closed subset in , and let 𝑟>0. Assume that 𝑆 is 𝑟-prox-regular. Then the following holds: 𝑥,with𝑑𝑆(𝑥)<𝑟,andall𝜁𝜕𝑃𝑑𝑆(𝑥)onehas𝜁,𝑥8𝑥𝑟𝑑𝑆𝑥(𝑥)𝑥2+𝑑𝑆𝑥𝑑𝑆(𝑥),𝑥with𝑑𝑆𝑥𝑟.(2.6)

The following proposition summarizes some important consequences of the uniform prox-regularity needed in the sequel of the paper (see [11, 13]).

Proposition 2.2. Let 𝑆 be a nonempty closed subset of and 𝑥𝑆. The following assertions hold.(1)𝜕𝑃𝑑𝑆(𝑥)=𝑁𝑃𝑆(𝑥)𝔹. (2)If 𝑆 is uniformly 𝑟-prox-regular, then 𝜕𝑃𝑑𝑆(𝑥) is a closed convex set in , and for any 𝑥 with 𝑑𝑆(𝑥)<𝑟, one has proj𝑆(𝑥).

Now, we recall some preliminaries concerning set-valued mappings.

Let 𝑇>0, 𝐼=[0,𝑇], and let 𝐾𝐼×𝐻 be a set-valued mapping. We will say that 𝐾 is Hausdorff-continuous (resp., Lipschitz with ratio (𝜆1,𝜆2)) if for any 𝑡𝐼 and 𝑥𝑋 one haslim𝑥𝑥𝐾(𝑡,𝑥),𝐾𝑡,𝑥=0(2.7) (resp., if for any 𝑥,𝑥𝑋 and 𝑡,𝑡𝐼 one has𝐾𝑡(𝑡,𝑥),𝐾,𝑥𝜆1𝑡𝑡+𝜆2𝑥𝑥.(2.8) Here denotes the Hausdorff distance relative to the norm associated with the Hilbert space defined by (𝐴,𝐵)=maxsup𝑎𝐴𝑑𝐵(𝑎),sup𝑏𝐵𝑑𝐴(𝑏).(2.9) Now, we give the following proposition. It proves the result of closedness of the proximal subdifferential of the distance function of prox-regular set. In [12], the authors proved the result when the set-valued mapping depends only on 𝑡𝐼. We adapt the proof to the case of set-valued mapping depending on two variables 𝑡 and 𝑥. For the completeness of the paper, we give the proof.

Proposition 2.3. Let 𝑟>0;𝐾𝐼×Ω be a Hausdorff-continuous set-valued mapping with uniformly 𝑟-prox regular values. For a given 0<𝛿<𝑟, the following holds:
“for any 𝑡𝐼 and 𝑧Ω,𝑥𝐾(𝑡,𝑧)+(𝑟𝛿)𝐵,𝑥𝑛𝑥,𝑡𝑛𝑡,𝑧𝑛𝑧(𝑡𝑛𝐼, 𝑥𝑛𝑖𝑠𝑛𝑜𝑡𝑛𝑒𝑐𝑒𝑠𝑠𝑎𝑟𝑦𝑖𝑛𝐾(𝑡𝑛,𝑧𝑛)) and 𝜉𝑛𝜉 with 𝜉𝑛𝜕𝑃𝑑𝐾(𝑡𝑛,𝑧𝑛)(𝑥𝑛); then 𝜉𝜕𝑃𝑑𝐾(𝑡,𝑧)(𝑥)”. Here “” denotes the weak convergence in .

Proof. Fix any 𝑧Ω,𝑡𝐼, and 𝑥𝐾(𝑡,𝑧)+(𝑟𝛿)𝐵. As 𝑡𝑛𝑡,𝑧𝑛𝑧, and 𝑥𝑛𝑥, one gets for 𝑛 sufficiently large 𝑡𝑛𝑡+(𝛿/4)𝐵 and 𝑥𝑛𝑥+(𝛿/4)𝐵. On the other hand, since 𝐾(𝑡,𝑧) is 𝑟-prox regular, one can choose (by Proposition 2.2) a point 𝑦𝐾(𝑡,𝑧) s.t. 𝑦𝑥=𝑑𝐾(𝑡,𝑧)(𝑥). So for every 𝑛 large enough, one can write ||𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛𝑑𝐾(𝑡,𝑧)𝑦||𝐾𝑡𝑛,𝑧𝑛,𝐾𝑡,𝑧+𝑥𝑛𝑦,(2.10) and hence the Hausdorff-continuity of 𝐾 yields for 𝑛 large enough 𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛𝛿4+𝑥𝑛𝑥+𝑥𝑦𝛿4+𝛿4+𝑟𝛿<𝑟.(2.11) For 𝑛 large enough and by Proposition 2.1, we have 𝜉𝑛,𝑢𝑥𝑛8𝑟𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛𝑢𝑥𝑛2+𝑑𝐾(𝑡𝑛,𝑧𝑛)(𝑢)𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛𝑢,𝑑𝐾(𝑡𝑛,𝑧𝑛)(𝑢)𝑟.(2.12) This inequality still holds for all 𝑢̀𝑥+𝛿𝐵 where ̀0<𝛿<𝛿/4 because 𝑑𝐾(𝑡𝑛,𝑧𝑛)(𝑢)𝑢𝑥+𝑥𝑥𝑛+𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛𝑢𝑥+𝑥𝑥𝑛+𝑑𝐾(𝑡𝑛,𝑧𝑛)𝑥𝑛̀𝛿𝛿+4𝛿+𝑟2<𝛿4+𝛿4𝛿+𝑟2=𝑟.(2.13) Consequently, by the continuity of the distance function with respect to (𝑡,𝑧,𝑥) (because of (2.7)), the inequality (2.12) gives, by letting 𝑛, 𝜉,𝑢𝑥8𝑟𝑑𝐾(𝑡,𝑧)𝑥𝑢𝑥2+𝑑𝐾(𝑡,𝑧)(𝑢)𝑑𝐾(𝑡,𝑧)𝑥𝑢̀𝑥+𝛿𝐵.(2.14) This ensures that 𝜉𝜕𝑃𝑑𝐾(𝑡,𝑧)(𝑥), and so the proof is complete.

3. Existence Results for Third-Order Nonconvex State-Dependent Sweeping Process with Perturbation

Throughout this section, will denote a separable Hilbert space. Let 𝑥0,𝑢0, 𝐴(𝜐0)𝐾(0,𝑢0), 𝜍,>0, 𝑇>0, 𝑈0, 𝑉0 be open neighborhoods of 𝑢0,𝑥0 (resp.) in such that 𝑥0+𝜍𝔹𝑈0,𝑢0+𝜍𝔹𝑉0, and 𝐾[0,𝜍/]×𝑐𝑙(𝑈0) be a Lipschitz set-valued mapping with ratio (𝜆1,𝜆2) taking nonempty closed uniformly 𝑟-prox regular values in . Assume that 𝐴 is a surjective bounded linear operator. Our aim is to prove the local existence of solution of (TSPMP), that is, there exist 𝑇>0 and Lipschitz mappings 𝑥[0,𝑇]𝑐𝑙(𝑉0),𝑢[0,𝑇]𝑐𝑙(𝑈0), and 𝜐[0,𝑇] such that𝑥(0)=𝑥0,𝑢(0)=𝑢0,𝜐(0)=𝜐0[];,𝐴(𝜐(𝑡))𝐾(𝑡,𝑢(𝑡)),a.e.on0,𝑇𝑥(𝑡)=𝑥0+𝑡0𝑢(𝑠)𝑑𝑠,𝑢(𝑡)=𝑢0+𝑡0[];𝜐(𝑠)𝑑𝑠,a.e.on0,𝑇𝐴(̇𝜐(𝑡))𝑁𝐾(𝑡,𝑢(𝑡))[].(𝐴(𝜐(𝑡)))+𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡)))+𝐺(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡)))a.e.0,𝑇(3.1) We begin by recalling the following lemma proved in [14] and needed in the proof of next theorem.

Lemma 3.1. Let (𝑋,𝑑𝑋) and (𝑌,𝑑𝑌) be two metric spaces, and let 𝑋𝑌 be a uniformly continuous mapping. Then for every sequence (𝜖𝑛)𝑛1 of positive numbers, there exists a strictly decreasing sequence of positive numbers (𝑒𝑛)𝑛1 converging to 0 such that(i)for any 𝑛2,1/(𝑒𝑛1) and (𝑒𝑛1)/𝑒𝑛 are integers 2;(ii)for any 𝑛1 and any 𝑥,𝑥𝑋, one has 𝑑𝑋𝑥,𝑥𝑒𝑛𝑑𝑌𝑥(𝑥),𝜖𝑛.(3.2)

We prove our main theorem in this section.

Theorem 3.2. Let 𝐹,𝐺[0,)×𝒞0×𝒞0×𝒞0 be two set-valued mappings, and let 𝜍>0 such that 𝑢0+𝜍𝔹𝑈0. Assume that the following assumptions are satisfied:(1)for all 𝑡[0,𝜍/] and 𝑢𝑐𝑙(𝑈0),𝐾(𝑡,𝑢)𝐴(𝜅1)𝑏𝔹, for some convex compact set 𝜅1 and some >0;(2)𝐹 is scalarly u.s.c. on [0,𝜍/]×𝑉0×𝑈0×Im𝐾 with nonempty convex weakly compact values;(3)𝐺 is uniformly continuous on [0,𝜍/]×𝛽𝔹𝒞0×𝛼𝔹𝒞0×𝑏𝔹𝒞0 into nonempty compact subset of , for 𝛼=𝑢0+𝜍 and 𝛽=𝑥0+𝛼(𝜍/) with 𝑥0+𝛼(𝜍/)𝔹𝑉0;(4)𝐹 and 𝐺 satisfy the linear growth condition, that is,𝐹(𝑡,𝑥,𝑢,𝜐)𝜌1(1+𝑥+𝑢+𝜐)𝔹,𝐺(𝑡,𝑥,𝑢,𝜐)𝜌2(1+𝑥+𝑢+𝜐)𝔹,(3.3) for all (𝑡,𝑥,𝑢,𝜐)[0,𝜍/]×𝑉0×𝑈0×Im𝐾 for some 𝜌1,𝜌20. Then for every 𝑇(0,𝜍/], there exist Lipschitz mappings 𝑥[0,𝑇]𝑐𝑙(𝑉0),𝑢[0,𝑇]𝑐𝑙(𝑈0), and 𝜐[0,𝑇] satisfying (TSPMP) with ̇𝑢(𝑡) and ̇𝜐(𝑡)(1/𝑏)(𝜆1+𝜆2+2(𝜌1+𝜌2)(1+𝛽+𝛼+𝑏)) a.e. on [0,𝑇].

Proof. We give the proof in four steps.
Step 1. Construction of the approximants.
Let 𝑇]0,𝜍/] and put 𝐼=[0,𝑇] and 𝜅=𝐼×𝛽𝔹×𝛼𝔹×𝑏𝔹. Then by the linear growth condition of 𝐹 and 𝐺 we have 𝐹(𝑡,𝑥,𝑢,𝜐)𝜌1(1+𝑥+𝑢+𝜐)𝜌1(1+𝛽+𝛼+𝑏)=𝜁1,𝐺(𝑡,𝑥,𝑢,𝜐)𝜌2(1+𝑥+𝑢+𝜐)𝜌2(1+𝛽+𝛼+𝑏)=𝜁2,(3.4) for all (𝑡,𝑥,𝑢,𝜐)𝜅(𝐼×𝑉0×𝑈0×Im𝐾) (since (0,𝑥0,𝑢0,𝐴(𝜐0))𝜅𝐼×𝑉0×𝑈0×Im𝐾).
Let 𝜖𝑛=1/2𝑛,(𝑛=1,2,3,). Then by the uniform continuity of 𝐺 and Lemma 3.1, there is a strictly decreasing sequence of positive numbers (𝑒𝑛) converging to 0 such that 𝑒𝑛1, 1/𝑒𝑛1 and 𝑒𝑛1/𝑒𝑛 are integers 2, and the following implication holds: 𝑡(𝑡,𝑥,𝑢,𝜐),𝑥,𝑢,𝜐𝜂𝑒𝑛𝐺𝑡(𝑡,𝑥,𝑢,𝜐),𝐺,𝑥,𝑢,𝜐𝜖𝑛,(3.5) for every (𝑡,𝑥,𝑢,𝜐),(𝑡,𝑥,𝑢,𝜐)𝜅 where 𝜂=(1+𝛼+3+𝜆1+(1+𝑇+𝜆2)+2(𝜁1+𝜁2)). Fix 𝑛0 so that 𝜆1+𝜆2+𝜁1+𝜁2𝑒𝑛0𝑟2.(3.6) For all 𝑛𝑛0, we consider the following partition of 𝐼. Without loss of generality we assume that 𝑇 is integer: 𝑃𝑛𝑡=𝑛,𝑖=𝑖𝑒𝑛𝑖=0,1,,𝜇𝑛=𝑇𝑒𝑛.(3.7)
Put 𝐼𝑛,𝑖=[𝑡𝑛,𝑖,𝑡𝑛,𝑖+1) for all 𝑖=0,,𝜇𝑛1 and 𝐼𝑛,𝜇𝑛={𝑇}. For every 𝑛𝑛0, we define the following approximating mappings on each interval 𝐼𝑛,𝑖 as 𝜐𝑛(𝑡)=𝜐𝑛,𝑖,𝑢𝑛(𝑡)=𝑢0+𝑡0𝜐𝑛(𝑠)𝑑𝑠,𝑥𝑛(𝑡)=𝑥0+𝑡0𝑢𝑛(𝑓𝑠)𝑑𝑠,𝑛(𝑡)=𝑓𝑛,𝑖𝑡𝐹𝑛,𝑖,𝑥𝑛𝑡𝑛,𝑖,𝑢𝑛𝑡𝑛,𝑖𝜐,𝐴𝑛,𝑖,𝑔𝑛(𝑡)=𝑔𝑛,𝑖𝑡𝐺𝑛,𝑖,𝑥𝑛𝑡𝑛,𝑖,𝑢𝑛𝑡𝑛,𝑖𝜐,𝐴𝑛,𝑖,(3.8) where 𝜐𝑛,0=𝜐0 and for all 𝑖=0,,𝜇𝑛1. Since 𝐴 is surjective, we can choose 𝜐𝑛,𝑖+1𝐻 such that 𝐴𝜐𝑛,𝑖+1=𝜔𝑛,𝑖+1,(3.9) where 𝜔𝑛,𝑖+1𝐴𝜐=Proj𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝑡,𝐾𝑛,𝑖+1,𝑢𝑛𝑡𝑛,𝑖+1.(3.10) This algorithm is well defined. Indeed, as 𝐴𝜐𝑛(𝑡)=𝜔𝑛,𝑖𝑡𝐾𝑛,𝑖,𝑢𝑛𝑡𝑛,𝑖𝜅𝐴1𝑏𝐵,(3.11) therefore, by (1) and (3.11), we get 𝑏𝜐𝑛,𝑖𝐴𝜐𝑛,𝑖𝑏,(3.12) so that 𝜐𝑛,𝑖.(3.13) Then by the Lipschitz property of 𝐾 and the relations (3.4), (3.6), (3.8), (3.10), and (3.13), we get 𝑑𝐾(𝑡𝑛,𝑖+1,𝑢𝑛,𝑖+1)𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝐾𝑡𝑛,𝑖+1,𝑢𝑛𝑡𝑛,𝑖+1𝑡,𝐾𝑛,𝑖,𝑢𝑛𝑡𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝜆1𝑡𝑛,𝑖+1𝑡𝑛,𝑖+𝜆2𝑢𝑛𝑡𝑛,𝑖+1𝑢𝑛𝑡𝑛,𝑖+𝑒𝑛𝜁1+𝜁2𝜆1𝑒𝑛+𝜆2𝑡𝑛,𝑖+1𝑡𝑛,𝑖𝜐𝑛,𝑖+𝑒𝑛𝜁1+𝜁2𝜆1+𝜆2+𝜁1+𝜁2𝑒𝑛𝜆1+𝜆2+𝜁1+𝜁2𝑒𝑛0𝑟2<𝑟.(3.14) Therefore, as 𝐾 has uniformly 𝑟-prox-regular values, by Proposition 2.2 one can choose a point 𝜔𝑛,𝑖+1𝐴𝜐=Proj𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝑡,𝐾𝑛,𝑖+1,𝑢𝑛𝑡𝑛,𝑖+1.(3.15) Define 𝜃𝑛𝐼𝐼 by 𝜃𝑛(0)=0, and 𝜃𝑛(𝑡)=𝑡𝑛,𝑖,𝑡𝐼𝑛,𝑖.(3.16) Then (3.11) becomes 𝐴𝜐𝑛𝜃(𝑡)𝐾𝑛(𝑡),𝑢𝑛𝜃𝑛(𝑡).(3.17) So that, all the mappings 𝑢𝑛 are Lipschitz with ratio , and they are also equibounded, with 𝑢𝑛(𝑡)𝑢0+𝑇. Observe that for all 𝑛𝑛0 and all 𝑡𝐼, one has 𝑢𝑛(𝑡)𝛼𝐵𝑈0,(3.18) because 𝑢𝑛(𝑡)=𝑢0+𝑡0𝜐𝑛(𝑠)𝑑𝑠𝑢0+𝑡𝔹𝑢0+𝜍𝔹𝛼𝔹𝑈0.(3.19) We note that 𝑥𝑛𝜃𝑛(𝑡)=𝑥𝑛𝑡𝑛,𝑖=𝑥0+𝑡𝑛,𝑖0𝑢𝑛(𝑠)𝑑𝑠𝑥0+𝑡𝑛,𝑖𝛼𝔹𝑥0𝜍+𝛼𝔹𝛽𝔹𝑉0.(3.20) Now we define the affine approximants 𝑧𝑛(𝑡)=𝜐𝑛,𝑖+𝑒𝑛1𝑡𝑡𝑛,𝑖𝜐𝑛,𝑖+1𝜐𝑛,𝑖,if𝑡𝐼𝑛,𝑖.(3.21) Observe that 𝐴𝑧𝑛𝜃𝑛𝜐(𝑡)=𝐴𝑛,𝑖𝜃𝐾𝑛(𝑡),𝑢𝑛𝜃𝑛(𝑡)𝑏𝔹.(3.22) Then by (3.4), (3.8), (3.18), (3.20), and (3.22), we can write 𝑓𝑛(𝑡)=𝑓𝑛,𝑖𝜃𝐹𝑛(𝑡),𝑥𝑛𝜃𝑛(𝑡),𝑢𝑛𝜃𝑛𝑧(𝑡),𝐴𝑛𝜃𝑛(𝑡)𝜁1𝑔𝔹,𝑛(𝑡)=𝑔𝑛,𝑖𝜃𝐺𝑛(𝑡),𝑥𝑛𝜃𝑛(𝑡),𝑢𝑛𝜃𝑛(𝑧𝑡),𝐴𝑛𝜃𝑛(𝑡)𝜁2𝔹.(3.23) Now, we check that the mappings 𝑧𝑛 are equi-Lipschitz with ratio (1/𝑏)(𝜆1+𝜆2+2(𝜁1+𝜁2)). Indeed, by (3.10) and (3.11), one has 𝜔𝑛,𝑖+1𝜔𝑛,𝑖𝜔𝑛,𝑖+1𝜔𝑛,𝑖𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖=𝜔𝑛,𝑖+1𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝑑𝐾(𝑡𝑛,𝑖+1,𝑢𝑛,𝑖+1)𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖+𝑒𝑛𝜁1+𝜁2𝜆1+𝜆2𝜁+21+𝜁2𝑒𝑛.(3.24) From the assumption (1), (3.24) becomes 𝜐𝑛,𝑖+1𝜐𝑛,𝑖1𝑏𝐴𝜐𝑛,𝑖+1𝜐𝑛,𝑖=1𝑏𝜔𝑛,𝑖+1𝜔𝑛,𝑖1𝑏𝜆1+𝜆2𝜁+21+𝜁2𝑒𝑛.(3.25) So for any 𝑡,𝑠𝐼𝑛,𝑖𝑧𝑛(𝑡)𝑧𝑛(𝑠)=𝑒𝑛1𝜐|𝑡𝑠|𝑛,𝑖+1𝜐𝑛,𝑖1𝑏𝜆1+𝜆2𝜁+21+𝜁2|𝑡𝑠|.(3.26) By addition on all the interval 𝐼, we obtain the Lipschitz property of 𝑧𝑛 on all 𝐼. Clearly, by the definition of 𝑧𝑛() and (3.25), we have 𝑧𝑛(𝑡)𝜐𝑛(𝑡)𝑒𝑛1||𝑡𝑡𝑛,𝑖||𝜐𝑛,𝑖+1𝜐𝑛,𝑖1𝑏𝜆1+𝜆2𝜁+21+𝜁2𝑒𝑛,(3.27) and so 𝑧𝑛𝜐𝑛0as𝑛.(3.28) Let us define 𝜌𝑛(𝑡)=𝑡𝑛,𝑖+1, for all 𝑡𝐼𝑛,𝑖. Then we have 𝐴𝑧𝑛𝜌𝑛𝜌(𝑡)𝐾𝑛(𝑡),𝑢𝑛𝜌𝑛(𝑡).(3.29) Coming back to the definition of 𝑧𝑛, one observes that for a.e. 𝑡𝐼, ̇𝑧𝑛(𝑡)=𝑒𝑛1𝜐𝑛,𝑖+1𝜐𝑛,𝑖.(3.30) So, for a.e. 𝑡𝐼, 𝑒𝑛𝐴̇𝑧𝑛𝑓(𝑡)𝑛(𝑡)+𝑔𝑛(𝑡)=𝜔𝑛,𝑖+1𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖=Proj𝐾(𝑡𝑛,𝑖+1,𝑢𝑛(𝑡𝑛,𝑖+1))𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖𝜔𝑛,𝑖+𝑒𝑛𝑓𝑛,𝑖+𝑔𝑛,𝑖.(3.31) Then, by properties of proximal normal cone, we have for a.e. 𝑡𝐼, 𝐴̇𝑧𝑛𝑓(𝑡)𝑛(𝑡)+𝑔𝑛𝐾𝑡(𝑡)𝑁𝑛,𝑖+1,𝑢𝑛𝑡𝑛,𝑖+1𝜐;𝐴𝑛,𝑖+1𝐾𝜌=𝑁𝑛(𝑡),𝑢𝑛𝜌𝑛(𝜐𝑡);𝐴𝑛𝜌𝑛(.𝑡)(3.32) On the other hand, by (3.25) and (3.30), we have 𝐴̇𝑧𝑛1(𝑡)𝐴𝑏𝜆1+𝜆2𝜁+21+𝜁2.(3.33) Put 𝛿=(𝐴/𝑏)(𝜆1+𝜆2)+((2𝐴/𝑏)+1)(𝜁1+𝜁2). Therefore, the relations (3.32) and (3.33) and Proposition 2.2 entail for a.e. 𝑡𝐼𝐴̇𝑧𝑛𝑓(𝑡)𝑛(𝑡)+𝑔𝑛(𝑡)𝛿𝜕𝑑𝐾(𝜌𝑛(𝑡),𝑢𝑛(𝜌𝑛(𝑡)))𝐴𝜐𝑛𝜌𝑛.(𝑡)(3.34)
Step 2. We will prove the uniform convergence of 𝑢𝑛 and 𝑧𝑛.
Since 𝑒𝑛1(𝑡𝑡𝑛,𝑖)1 for all 𝑡𝐼𝑛,𝑖 and 𝜐𝑛,𝑖+1,𝜐𝑛,𝑖𝜅1 and 𝐴1𝜅1 is a convex set in , one gets 𝑧𝑛(𝑡)𝜅1 for all 𝑡𝐼 so that, for every 𝑡𝐼, the set {𝑧𝑛(𝑡)𝑛𝑛0} is relatively compact. By using Arzela-Ascoli theorem, there exists a Lipschitz mapping 𝜐𝐼 with ratio (1/𝑏)(𝜆1+𝜆2+2(𝜁1+𝜁2)) such that, (𝑧𝑛) converges uniformly to 𝜐 on 𝐼; (̇𝑧𝑛) weakly star converges to ̇𝜐 in 𝐿(𝐼,).
Now, we define the Lipschitz mappings 𝑢𝐼 and 𝑥𝐼 as follows: 𝑢(𝑡)=𝑢0+𝑡0𝜐(𝑠)𝑑𝑠,𝑡𝐼,𝑥(𝑡)=𝑥0+𝑡0𝑢(𝑠)𝑑𝑠,𝑡𝐼.(3.35) By (3.28), we have 𝑣𝑛𝑣𝜐𝑇𝑛𝑧𝑛𝑧+𝑇𝑛𝑣0as𝑛,(3.36) and so 𝑢𝑛(=𝑡)𝑢(𝑡)𝑡0𝜐𝑛(𝜐𝑠)𝜐(𝑠)𝑑𝑠𝑇𝑛𝜐.(3.37) This ensures that 𝑥𝑛𝑥𝑢𝑇𝑛𝑢0as𝑛.(3.38) Thus, 𝑢𝑛,𝑥𝑛, and 𝑣𝑛 uniformly converge to 𝑢,𝑥, and 𝑣, respectively.
Step 3 (relative strong compactness of (𝑔𝑛)). The points (𝑔𝑛,𝑖) defining the step function 𝑔𝑛 were chosen arbitrarily in our construction. Nevertheless, by using the uniform continuity of the set-valued mapping 𝐺 over 𝜅 and the techniques of [14] (see also [15, 16]), the sequence 𝑔𝑛 can be constructed relatively strongly compact for the uniform convergence in the space of bounded functions. Therefore, there exists a bounded mapping 𝑔𝐼𝐻 such that 𝑔𝑛𝑔0.Step 4 (existence of solutions). Since |𝜌𝑛(𝑡)𝑡|𝑒𝑛 on [0,𝑇], then 𝜌𝑛(𝑡)𝑡, as well as 𝜃𝑛(𝑡)𝑡. Also (𝑥𝑛𝜃𝑛),(𝑢𝑛𝜃𝑛), and (𝜐𝑛𝜃𝑛) converge uniformly to 𝑥, 𝑢, and 𝜐, respectively. Then by continuity of 𝐺 on 𝜅, the closedness of the set 𝐺(𝑡,𝑥(𝑡),𝑢(𝑡), 𝐴(𝜐(𝑡))), and by (3.23), we obtain 𝑔(𝑡)𝐺(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))) a.e. on 𝐼.
Since 𝐴(𝜐𝑛(𝜃𝑛(𝑡))𝐾(𝜃𝑛(𝑡),𝑢𝑛(𝜃𝑛(𝑡))) and so by the closedness and the continuity of 𝐾 and the continuity of 𝐴, we have 𝐴(𝜐(𝑡))𝐾(𝑡,𝑢(𝑡)) a.e. on 𝐼. By (3.23), the sequence 𝑓𝑛 is bounded in 𝐿(𝐼,) with separable; so 𝑓𝑛 is relatively sequentially 𝜎(𝐿(𝐼,),𝐿1(𝐼,))-compact in 𝐿(𝐼,), because 𝐿(𝐼,) is the dual of the separable Banach space 𝐿1(𝐼,). Therefore, by integrating, we have Ω𝑥,𝑓(𝑡)𝑑𝑡=lim𝑛Ω𝑥,𝑓𝑛(𝑡)𝑑𝑡limsup𝑛Ω𝜎𝜃𝑥,𝐹𝑛(𝑡),𝑥𝑛𝜃𝑛(𝑡),𝑢𝑛𝜃𝑛𝑧(𝑡),𝐴𝑛𝜃𝑛(𝑡)𝑑𝑡Ωlimsup𝑛𝜎𝜃𝑥,𝐹𝑛(𝑡),𝑥𝑛𝜃𝑛(𝑡),𝑢𝑛𝜃𝑛𝑧(𝑡),𝐴𝑛𝜃𝑛(𝑡)𝑑𝑡Ω𝜎(𝑥,𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))))𝑑𝑡,(3.39) for every measurable set Ω in 𝐼 and every 𝑥. By assumption, the set-valued mapping 𝑡𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))) is measurable with convex weakly compact values, which ensures the last inequality 𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))),a.e.on𝐼.(3.40) Using the same technique, we get 𝐴(̇𝜐(𝑡))𝑓(𝑡)𝑔(𝑡)𝛿𝜕𝑑𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡))).(3.41) Indeed, for every measurable set Ω in 𝐼 and every 𝜉, we have Ω𝜉,𝐴(̇𝜐(𝑡))𝑓(𝑡)𝑔(𝑡)𝑑𝑡=lim𝑛Ω𝜉,𝐴̇𝑧𝑛(𝑡)𝑓𝑛(𝑡)𝑔𝑛(𝑡)𝑑𝑡limsup𝑛Ω𝜎𝜉,𝛿𝜕𝑑𝐾(𝜌𝑛(𝑡),𝑢𝑛(𝜌𝑛(𝑡)))𝐴𝜐𝑛𝜌𝑛(𝑡)𝑑𝑡Ωlimsup𝑛𝜎𝜉,𝛿𝜕𝑑𝐾(𝜌𝑛(𝑡),𝑢𝑛(𝜌𝑛(𝑡)))𝐴𝜐𝑛𝜌𝑛(𝑡)𝑑𝑡Ω𝜎𝜉,𝛿𝜕𝑑𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))𝑑𝑡.(3.42) Thus, as the set-valued mapping 𝑡𝛿𝜕𝑑𝐾(𝑡,𝑢(𝑡))(𝐴(𝑣(𝑡))) is measurable with convex weakly compact values (see [17]), it follows that 𝐴(̇𝜐(𝑡))𝑓(𝑡)𝑔(𝑡)𝛿𝜕𝑑𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡))),(3.43) and since 𝐴(𝜐(𝑡))𝐾(𝑡,𝑢(𝑡)) a.e. on 𝐼, we get for a.e. 𝑡𝐼, 𝐴(̇𝜐(𝑡))𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝑓(𝑡)+𝑔(𝑡)𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡)))+𝐺(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))).(3.44) Thus the proof of the theorem is complete.

We deduce from our main theorem an existence result for a second-order nonconvex differential inclusion.

Corollary 3.3. Let 𝑥0,𝑢0, 𝜐0𝐾(0,𝑢0),𝜍,>0,𝑈0,𝑉0 be open neighborhoods of 𝑢0,𝑥0 (resp.) in such that 𝑥0+𝜍𝔹𝑈0, 𝑢0+𝜍𝔹𝑉0,and𝐾[0,𝜍/]×𝑐𝑙(𝑈0) is a Lipschitz set-valued mapping with ratio (𝜆1,𝜆2) taking nonempty closed uniformly 𝑟-prox regular values in . Assume that 𝐾(𝑡,𝑢)𝜅𝔹,forall(𝑡,𝑥)[0,𝜍/]×𝑐𝑙(𝑈0), for some convex compact set 𝜅. Then for any 𝑡(0,𝜍/], there exists a Lipschitz solution of the second-order differential inclusion ̈𝑥(𝑡)𝐾(𝑡,̇𝑥(𝑡))a.e.on𝐼.(3.45)

Proof. Take 𝐹=𝐺=0,𝐴=𝐼𝑑 in Theorem 3.2. Then there is a Lipschitz solution 𝑥[0,𝑇] to the cauchy problem for the third-order differential inclusion 𝑥(3)(𝑡)𝑁𝐾(𝑡,̇𝑥(𝑡))[];(̈𝑥(𝑡))a.e.0,𝑇𝑥(0)=𝑥0,̇𝑥(0)=𝑢0[],,̈𝑥(𝑡)𝐾(𝑡,̇𝑥(𝑡))a.e.on0,𝑇(3.46) which proves that ̈𝑥(𝑡)𝐾(𝑡,̇𝑥(𝑡)) has a solution almost everywhere on [0,𝑇].

Remark 3.4. The existence result proved in Theorem 3.2 cannot be covered by the recent existence result for third-order differential inclusions established in [18] in finite-dimensional case and extended in [19] in Hilbert spaces. Indeed, The right-hand side in Theorem 3.2 contains the normal cone, which cannot be bounded nor u.s.c. as a set-valued mapping. These two assumptions are essential in the proof of the results in [18, 19].

4. Solution Set

Throughout this section, let 𝑇>0and𝐼=[0,𝑇], and let 𝐹𝐼××× be a set valued mapping, Ω1,Ω2 open subsets in , and 𝐾𝐼×𝑐𝑙(Ω) a Lipschitz set-valued mapping with ratio (𝜆1,𝜆2) taking nonempty closed uniformly 𝑟-prox regular values in . Let 𝑥0Ω1,𝑢0Ω2,𝐴(𝜐0)𝐾(0,𝑢0) with 𝑢0+𝑇Ω2. We denote 𝑆(𝑥0,𝑢0,𝜐0) the set of all triple (𝑥,𝑢,𝜐) of Lipschitz mappings 𝑥,𝑢,𝜐𝐼𝐻 such that𝐴(̇𝜐(𝑡))𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))),a.e.on𝐼,𝑥(𝑡)=𝑥0+𝑡0𝑢(𝑠)𝑑𝑠,𝑢(𝑡)=𝑢0+𝑡0𝜐(𝑠)𝑑𝑠,𝑥(0)=𝑥0,𝑢(0)=𝑢0,𝜐(0)=𝜐0,𝐴(𝜐(𝑡))𝐾(𝑡,𝑢(𝑡)),a.e.on𝐼.(4.1)

Proposition 4.1. Assume that the hypothesis on 𝐹 and 𝐾 in Theorem 3.2 are satisfied and let 𝐺=0. Then the graph of the set-valued mapping 𝑆 is closed in Ω1×Ω2×Im𝐾×𝐶(𝐼;××).

Proof. Let ((𝑥𝑛0,𝑢𝑛0,𝜐𝑛0))𝑛Ω1×Ω2×Im𝐾 and ((𝑥𝑛,𝑢𝑛,𝜐𝑛))𝑛𝐶(𝐼;××) with (𝑥𝑛,𝑢𝑛,𝜐𝑛)𝑆(𝑥𝑛0,𝑢𝑛0,𝜐𝑛0) such that ((𝑥𝑛0,𝑢𝑛0,𝜐𝑛0))𝑛 uniformly converges to some (𝑥0,𝑢0,𝜐0)Ω1×Ω2×Im𝐾 and ((𝑥𝑛,𝑢𝑛,𝜐𝑛))𝑛 uniformly converges to some (𝑥,𝑢,𝜐)𝐶(𝐼;××). We have to show that (𝑥,𝑢,𝜐)𝑆(𝑥0,𝑢0,𝜐0). First, observe that for 𝑛 sufficiently large, 𝑢𝑛0𝑢0+𝜍𝔹𝑈0 and 𝑥𝑛0𝑥0+𝛼(𝜍/)𝔹𝑉0, where 𝛼 is given as in the proof of Theorem 3.2. Now, it is not difficult to check that the continuity of 𝐾, the continuity of 𝐴, the uniform convergence of both sequences ((𝑥𝑛0,𝑢𝑛0,𝜐𝑛0))𝑛 and ((𝑥𝑛,𝑢𝑛,𝜐𝑛))𝑛, and 𝐴(𝜐𝑛(𝑡))𝐾(𝑡,𝑢𝑛(𝑡)) for almost all 𝑡𝐼 imply that 𝐴(𝜐(𝑡))𝐾(𝑡,𝑢(𝑡)) for almost all 𝑡𝐼. On the other hand, we have 𝑢(𝑡)=lim𝑛𝑢𝑛(𝑡)=lim𝑛𝑢𝑛0+𝑡0𝜐𝑛(𝑠)𝑑𝑠=𝑢0+𝑡0𝜐(𝑠)𝑑𝑠,t𝐼,𝑥(𝑡)=lim𝑛𝑥𝑛(𝑡)=lim𝑛𝑥0+𝑡0𝑢𝑛(𝑠)𝑑𝑠=𝑥0+𝑡0𝑢(𝑠)𝑑𝑠,𝑡𝐼.(4.2) It remains then to show that 𝐴(̇𝜐(𝑡))𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))),a.e.on𝐼.(4.3) For every 𝑛, one has 𝐴(̇𝜐𝑛(𝑡))𝑁𝐾(𝑡,𝑢𝑛(𝑡))(𝐴(𝜐𝑛(𝑡)))+𝐹(𝑡,𝑥𝑛(𝑡),𝑢𝑛(𝑡),𝐴(𝜐𝑛(𝑡))),a.e.on𝐼.(4.4) Then for every 𝑛, there exist a measurable selection 𝑓𝑛 such that 𝑓𝑛(𝑡)𝐹(𝑡,𝑥𝑛(𝑡),𝑢𝑛(𝑡),𝐴(𝜐𝑛(𝑡))),𝐴(̇𝜐𝑛(𝑡))𝑓𝑛(𝑡)𝑁𝐾(𝑡,𝑢𝑛(𝑡))(𝐴(𝜐𝑛(𝑡))),(4.5) for a.e. 𝑡𝐼. By Theorem 3.2, one has for 𝑛 sufficiently large ̇𝜐𝑛1(𝑡)𝑏𝜆1+𝜆2𝜌+21+𝜌2𝜁,(4.6) where 𝜁=1+𝑥0+(1+𝑇)(𝑢0+𝑇)+𝐴. By (5) in Theorem 3.2, we have 𝑓𝑛(𝑡)𝜌1𝑥1+0+𝑢(1+𝑇)0+𝑇+𝐴.(4.7) Therefore, we get the 𝜎(𝐿(𝐼;),𝐿1(𝐼;))-convergence of subsequences of both ̇𝜐𝑛 and 𝑓𝑛 to ̇𝜐 and 𝑓, respectively, in 𝐿(𝐼;). Using the same techniques in the proof of Theorem 3.2, we can prove that 𝑓(𝑡)𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))) a.e. 𝑡𝐼. Now taking 𝛿=(𝐴/𝑏)(𝜆1+𝜆2+2(𝜌1+𝜌2)𝜁)+2(𝜌1+𝜌2)𝜁 and using Proposition 2.2 yield 𝐴(̇𝜐𝑛(𝑡))𝑓𝑛(𝑡)𝛿𝜕𝑑𝐾(𝑡,𝑢𝑛(𝑡)))𝐴𝜐𝑛.(𝑡)(4.8) Once again, we use the techniques employed in the proof of Theorem 3.2 to show that for a.e. 𝑡𝐼𝐴(̇𝜐(𝑡))+𝑓(𝑡)𝛿𝜕𝑑𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡))).(4.9) Thus we get for a.e. 𝑡𝐼𝐴(̇𝜐(𝑡))𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝑓(𝑡)𝑁𝐾(𝑡,𝑢(𝑡))(𝐴(𝜐(𝑡)))+𝐹(𝑡,𝑥(𝑡),𝑢(𝑡),𝐴(𝜐(𝑡))),(4.10) and so the proof is complete.

5. Applications to Differential Variational Inequalities

In this section, we are interested with an application of the main result proved in Theorem 3.2 to differential variational inequalities (DVI): given a convex compact set 𝐷 in , three points 𝑥0,𝑢0,𝑣0𝐷+𝑓(0,𝑢0) with Λ(𝑣0𝑓(0,𝑢0))0:[]Find𝑇>0andaLipschitzmapping𝑥0,𝑇suchthat(1)𝑥(0)=𝑥0,̇𝑥(0)=𝑢0,̈𝑥(0)=𝑣0;[];𝑥(2)Λ(̈𝑥(𝑡)𝑓(𝑡,̇𝑥(𝑡)))0,̈𝑥(𝑡)𝐷+𝑓(𝑡,̇𝑥(𝑡)),a.e.on0,𝑇(3)𝑤𝐷+𝑓(𝑡,̇𝑥(𝑡))withΛ(𝑤𝑓(𝑡,̇𝑥(𝑡)))0,wehave(3)(𝑡)𝛼𝑥(𝑡),𝑤̈𝑥(𝑡)𝑎(𝑥(𝑡)+̇𝑥(𝑡)̈𝑥(𝑡),𝑤̈𝑥(𝑡))+𝜌𝑤̈𝑥(𝑡)2[],,a.e.on0,𝑇(5.1) where 𝛼>0, Λ is inf-compact and lower-𝐶2 function, 𝑎(,) is a real bilinear, symmetric, bounded, and elliptic form on ×, and 𝑓𝐼× is a Lipschitz function. Let 𝐴 be a linear and bounded operator on associated with 𝑎(,), that is, 𝑎(𝑢,𝑣)=𝐴𝑢,𝑣,forall𝑢,𝑣. We use Theorem 3.2 to prove that (DVI) has at least one Lipschitz solution.

Proposition 5.1. Assume that is a separable Hilbert space and inf{𝜉𝜉𝜕Λ(𝑥)with𝑥𝐷}0.(5.2) Then (DVI) has at least one Lipschitz solution.

Proof. Let 𝑆={𝑥𝐷Λ(𝑥)0} and define 𝐾(𝑡,𝑦)=𝑆+𝑓(𝑡,𝑦), for all (𝑡,𝑦)×. Since Λ is inf-compact, the set 𝑆 is compact in and so 𝐾 has compact values. Also, the lower 𝐶2 property of the function Λ and the assumption inf{𝜉𝜉𝜕Λ(𝑥)with𝑥𝐷}0 ensure by Theorem  3.3 in [20] the uniform prox regularity for some 𝑟>0, and so 𝐾 has uniform 𝑟-prox-regular values. The Lipschitz behavior of the set-valued mapping 𝐾 is inherited from the function 𝑓. Now, we use the definition of proximal normal cones for uniform prox-regular sets to rewrite (DVI) in the form of (TSPMP) as follows: 𝑥(3)(𝑡)𝐴(𝑥(𝑡)+̇𝑥(𝑡)̈𝑥(𝑡))𝛼𝑥(𝑡)𝑁𝐾(𝑡,̇𝑥(𝑡))[],(̈𝑥(𝑡)),a.e.on0,𝑇𝑥(0)=𝑥0,̇𝑥(0)=𝑢0,̈𝑥(0)=𝑣0[].,̈𝑥(𝑡)𝐾(𝑡,̇𝑥(𝑡))a.e.on0,𝑇(5.3) Let, now, 𝐺(𝑡,𝑥,𝑦,𝑧)={𝐴(𝑥+𝑦𝑧)𝛼𝑥}, for all (𝑡,𝑥,𝑦,𝑧)×××. Clearly 𝐺 is uniformly continuous (since 𝐴 is a bounded linear operator) with compact values. Also 𝐺 satisfies the linear growth condition with 𝜌2=𝛼+𝐴. Indeed, 𝐺(𝑡,𝑥,𝑦,𝑧)𝛼𝑥+𝐴(𝑥+𝑦𝑧)(𝛼+𝐴)(1+𝑥+𝑦+𝑧). Consequently, all the assumptions of Theorem 3.2 are satisfied, and so we have the existence of a Lipschitz solution of (DVI).

Acknowledgment

This project was supported by the Research Center, College of Science, King Saud University.