ROCKY MOUNTAIN JOURNAL

OF MATHEMATICS
Volume 52 (2022), No. 6, 2257–2272

INFINITE INTERVAL PROBLEMS FOR

HILFER FRACTIONAL EVOLUTION EQUATIONS
WITH ALMOST SECTORIAL OPERATORS

M IAN Z HOU , YONG L IANG AND YONG Z HOU

We investigate the Cauchy problem on an infinite interval for the fractional evolution equation with Hilfer
fractional derivative, which is a generalization of both Riemann–Liouville and Caputo fractional deriva-
tives. Our methods are based on the generalized Ascoli–Arzelà theorem, Schauder’s fixed point theorem,
the Wright function and Kuratowski’s measure of noncompactness. We obtain sufficient conditions of the
existence for global mild solutions and attractive solutions when the semigroup associated with an almost
sectorial operator is compact as well as noncompact. Two examples are provided to illustrate the results.

1. Introduction

Fractional calculus is considered as a generalization of classical calculus. The order of fractional derivative
can be an arbitrary (noninteger) positive real number or even complex number. In the past two decades,
fractional calculus has been a research focus and attracted the attention of many researchers all over the
world. On the one hand, it is due to the further development of fractional calculus theory itself. More
than that, fractional calculus is more and more widely used in various disciplines, especially in fluid
mechanics, physics, signal processing, materials science, electrochemistry, biology and so on.
In recent years, fractional differential equations have been widely used in the mathematical modeling
of real-world phenomena. These applications have motivated many researchers in the field of differential
equations to investigate fractional differential equations with different fractional derivatives; see the
monographs [1; 7; 10; 15; 18; 19].
The main motivation of studying fractional evolution equation comes from two aspects. One is
that many mathematical models in physics and fluid mechanics are characterized by fractional partial
differential equations. Secondly, many types of fractional partial differential equations, such as fractional
diffusion equations, wave equations, Navier–Stokes equations, Rayleigh–Stokes equations, Fokker–
Planck equations, Schrödinger equations, and so on, can be abstracted as fractional evolution equations
[20; 23; 22]. Therefore, the study of fractional evolution equations is of great significance both in terms
of theory and practical application.

Yong Zhou is the corresponding author. The research was supported by the Macau Science and Technology Development Fund
(Grant No. 0092/2022/A).
2020 AMS Mathematics subject classification: 26A33, 34A08, 35R11.
Keywords and phrases: fractional evolution equations, Hilfer derivative, almost sectorial operator, infinite interval.
Received by the editors on February 17, 2022, and in revised form on February 26, 2022.

2257
DOI: 10.1216 / rmj.2022.52.2257 © Rocky Mountain Mathematics Consortium
2258 MIAN ZHOU, YONG LIANG AND YONG ZHOU

The well-posedness of fractional evolution equations is an important research topic of evolution

equations. However, it seems that there are few works concerned with the existence of fractional evolution
equations on infinite intervals. Almost all of these results involve the existence of solutions for fractional
evolution equations on a finite interval [0, a], where a ∈ (0, ∞) (see [3; 8; 9; 14; 16; 17]).
Consider the Cauchy problem of fractional evolution equations on infinite interval
ν,µ

 H D0+ x(t) = Ax(t) + f (t, x(t)), t ∈ (0, ∞),
(1)
 I (1−ν)(1−µ) x(0) = x ,
0+ 0

ν,µ (1−ν)(1−µ)
where H D0+ is the Hilfer fractional derivative of order 0 < µ < 1 and type 0 ≤ ν ≤ 1, I0+ is the
Riemann–Liouville integral of order (1 − ν)(1 − µ), A is an almost sectorial operator in Banach space X ,
and f : [0, ∞) × X → X is a function which is defined later.
The Hilfer fractional derivative is a natural generalization of Riemann–Liouville derivative and Caputo
derivative [7]. It is obvious that fractional differential equations with Hilfer derivatives include fractional
differential equations with Riemann–Liouville derivative or Caputo derivative as special cases.
In this paper, we prove existence theorems of global mild solutions for the infinite interval problem (1)
for the cases that the semigroup associated with the almost sectorial operator is compact as well as
noncompact. In particular, we do not need to assume that f (t, · ) satisfies the Lipschitz condition. Our
methods in this paper are based on the generalized Ascoli–Arzelà theorem, Schauder’s fixed point theorem
and Kuratowski’s measure of noncompactness.

2. Preliminaries

We first introduce some notations and definitions about almost sectorial operators, fractional calculus and
Kuratowski’s measure of noncompactness. For more details, we refer to [7; 10; 5; 14; 18].
Assume that X is a Banach space with the norm | · |. Let R = (−∞, ∞) and J be an infinite interval
of R. By C(J, X ) we denote the space of all continuous functions from J to X . We denote by L (X )
the space of all bounded linear operators from X to X with the usual operator norm ∥ · ∥L (X ) . Denote
by D(A) the domain of A, by σ (A) its spectrum, while ρ(A) := C − σ (A) is the resolvent set of A.
Let G 0λ = {z ∈ C\{0} : | arg z| < λ} be the open sector for 0 < λ < π, and G λ be its closure, i.e.,
G λ = {z ∈ C\{0} : | arg z| ≤ λ} ∪ {0}.

Definition 1. Let 0 < k < 1 and 0 < ω < π2 . We denote 2−k

ω (X ) as a family of all closed linear operators
A : D(A) ⊂ X → X such that

(i) σ (A) ⊂ G ω = {z ∈ C \ {0} : | arg z| ≤ ω} ∪ {0} and

(ii) for any λ ∈ (ω, π), there exists a Cλ such that

∥R(z; A)∥L (X ) ≤ Cλ |z|−k for all z ∈ C \ G λ ,

where R(z; A) = (z I − A)−1 , z ∈ ρ(A) is the resolvent operator of A. The linear operator A will be
called an almost sectorial operator on X if A ∈ 2−k
ω (X ).
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2259

We denote the semigroup associated with A by {T (t)}t≥0 . For t ∈ G 0π −ω

2
Z
1
T (t) = e−t z (A) = e−t z R(z; A) dz,
2πi 0ρ
where the integral contour 0ρ = {R+ eiρ } {R+ e−iρ } is oriented counterclockwise and ω < ρ < ν <
S
π
2 − |arg t| forms an analytic semigroup of growth order 1 − k. Throughout this paper, we have assumed
that T (t) is continuous in the uniform operator topology for t > 0.
Lemma 2 (see [18]). Assume that 0 < k < 1 and 0 < ω < π2 . Set A ∈ 2−k
ω (X ). Then

(i) T (s + t) = T (s)T (t) for any s, t ∈ G 0π −ω ;

2

(ii) there exists a constant C0 > 0 such that ∥T (t)∥L (X ) ≤ C0 t k−1 for any t > 0.
Definition 3 (see [10]). The fractional integral of order µ for a function x : [0, ∞) → R is defined as
Z t
µ 1
I0+ x(t) = (t − s)µ−1 x(s) ds, µ > 0, t > 0,
0(µ) 0
provided the right side is point-wise defined on [0, ∞), where 0(·) is the gamma function.
Definition 4 (Hilfer fractional derivative, see [7]). Let 0 < µ < 1 and 0 ≤ ν ≤ 1. The Hilfer fractional
derivative of order µ and type ν for a function x : [0, ∞) → R is defined as
H ν,µ d (1−ν)(1−µ)
ν(1−µ)
D0+ x(t) = I0+ I x(t).
dt 0+
Remark 5. (i) In particular, when ν = 0, 0 < µ < 1, then
H 0,µ d 1−µ µ
D0+ x(t) = I0+ x(t) =: LD 0+ x(t),
dt
µ
where LD 0+ is Riemann–Liouville derivative.
(ii) When ν = 1, 0 < µ < 1, we have
H 1,µ 1−µ d µ
D0+ x(t) = I0+ x(t) =: CD0+ x(t),
dt
µ
where CD 0+ is Caputo derivative.
Let D be a nonempty subset of X . The Kuratowski’s measure of noncompactness α is said to be
n n o
α(D) = inf d > 0 : D ⊂ M j and diam(M j ) ≤ d ,
S
j=1

where the diameter of M j is given by diam(M j ) = sup{|x − y| : x, y ∈ M j }, j = 1, . . . , n.

Lemma 6 [12]. Let {u n (t)}∞
n=1 : [0, ∞) → X be a continuous function family. If there exists ξ ∈ L [0, ∞)
1

such that
|u n (t)| ≤ ξ(t), t ∈ [0, ∞), n = 1, 2, . . . ,
then α({u n (t)}∞
n=1 ) is integrable on [0, ∞), and
nZ t o Z t
α u n (s) ds : n = 1, 2, . . . ≤ 2 α({u n (s) : n = 1, 2, . . .}) ds.
0 0
2260 MIAN ZHOU, YONG LIANG AND YONG ZHOU

Definition 7 [13]. Define the Wright function Wµ (θ ) by

∞
X (−θ )n−1
Wµ (θ ) = , 0 < µ < 1, θ ∈ C,
(n − 1)!0(1 − µn)
n=1

with the following property:

∞
0(1 + δ)
Z
θ δ Wµ (θ ) dθ = for δ ≥ 0.
0 0(1 + µδ)
Lemma 8. The problem (1) is equivalent to the integral equation
Z t
x0 (ν−1)(1−µ) 1
(2) x(t) = t + (t − s)µ−1 [Ax(s) + f (s, x(s))] ds, t ∈ (0, ∞).
0(ν(1 − µ) + µ) 0(µ) 0
Proof. This proof is similar to that in [2], so we omit it. □
Lemma 9. Assume that x(t) satisfies integral equation (2). Then
Z t
(3) x(t) = Pν,µ (t)x0 + Q µ (t − s) f (s, x(s)) ds, t ∈ (0, ∞),
0

where
Z ∞
ν(1−µ) µ−1
Pν,µ (t) = I0+ Q µ (t), Q µ (t) = t K µ (t), and K µ (t) = µθ Wµ (θ )T (t µ θ ) dθ.
0

Proof. This proof is similar to that in [3], so we omit it. □

In view of Lemma 8, we have the following definition.
Definition 10. If x ∈ C((0, ∞), X ) satisfies
Z t
x(t) = Pν,µ (t)x0 + Q µ (t − s) f (s, x(s)) ds, t ∈ (0, ∞),
0

then x(t) is called a mild solution of the Cauchy problem (1).

Definition 11. The mild solution x(t) of the Cauchy problem (1) is called attractive if x(t) → 0 as t → ∞.
Lemma 12 [3; 18]. If {T (t)}t>0 is a compact operator, then {Pν,µ (t)}t>0 and {K µ (t)}t.0 are also compact
operators.
Lemma 13 [8]. For any fixed t > 0, K µ (t), Q µ (t) and Pν,µ (t) are linear operators, and for any x ∈ X
|K µ (t)x| ≤ L 1 t µ(k−1) |x|, |Q µ (t)x| ≤ L 1 t µk−1 |x|, and |Pν,µ (t)x| ≤ L 2 t −1+ν−µν+µk |x|,
where
µC0 0(1 + k) L 1 0(µk)
L1 = and L2 = .
0(1 + µk) 0(ν(1 − µ) + µk)
Lemma 14 [8]. {K µ (t)}t>0 , {Q µ (t)}t>0 and {Pν,µ (t)}t>0 are strongly continuous, that is, for any x ∈ X
and t ′′ > t ′ > 0,
|K µ (t ′ )x −K µ (t ′′ )x| → 0, |Q µ (t ′ )x −Q µ (t ′′ )x| → 0 and |Pν,µ (t ′ )x − Pν,µ (t ′′ )x| → 0, as t ′′ → t ′ .
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2261

Let
E = u ∈ C([0, ∞), X ) : lim h(t)|u(t)| = 0 .

t→∞

Clearly, (E, ∥ · ∥) is a Banach space with the norm ∥u∥ = supt∈[0,∞) h(t)|u(t)| < ∞.
In the following, we state the generalized Ascoli–Arzelà theorem.
Lemma 15 [21]. The set 3 ⊂ E is relatively compact if and only if there is a function h ∈C([0, ∞), (0, ∞))
such that the following conditions hold:
(a) For any b > 0, the set V = {v : v(t) = h(t)x(t), x ∈ 3} is equicontinuous on [0, b].
(b) limt→∞ h(t)|x(t)| = 0 uniformly for x ∈ 3.
(c) For any t ∈ [0, ∞), V (t) = {v(t) : v(t) = h(t)x(t), x ∈ 3} is relatively compact in X .
Remark 16. Lemma 15 was given in [4] when h(t) := e−t and in [11] when h(t) := 1/(1 + t), t ∈ (0, ∞).

3. Some lemmas

First, we introduce the following hypotheses:

(H1) f (t, · ) is Lebesgue measurable with respect to t on [0, ∞). f ( ·, x) is continuous with respect
to x on X .
(H2) There exists a function m : (0, ∞) → (0, ∞) such that
µk
I0+ m(t) ∈ C((0, ∞), (0, ∞)), | f (t, x)| ≤ m(t) for all x ∈ X, t ∈ (0, ∞).
(H3) There exists a function h ∈ C([0, ∞), (0, ∞)) such that limt→∞ h(t) = 0 and
µk µk
lim t 1−ν+µν−µk h(t)I0+ m(t) = 0, lim t 1−ν+µν−µk h(t)I0+ m(t) = 0.
t→0+ t→∞

In this paper, we let

Cµ ((0, ∞), X )
= x ∈ C([0, ∞), X ) : lim t 1−ν+µν−µk x(t) exists and is finite, and lim h(t)t 1−ν+µν−µk |x(t)| = 0 ,

t→0+ t→∞

with the norm

∥x∥µ = sup h(t)t 1−ν+µν−µk |x(t)|.
t∈[0,∞)

Then (Cµ ((0, ∞), X ), ∥ · ∥µ ) is a Banach space.

For any x ∈ Cµ ((0, ∞), X ), define an operator S by
(Sx)(t) = (S1 x)(t) + (S2 x)(t),
where Z t
(S1 x)(t) = Pν,µ (t)x0 , (S2 x)(t) = Q µ (t − s) f (s, x(s)) ds for t ∈ (0, ∞).
0
Clearly, the problem (1) has a mild solution x ∈ Cµ ((0, ∞), X ) if and only if S has a fixed point
x ∈ Cµ ((0, ∞), X ).
For any u ∈ E, t ∈ (0, ∞), let x(t) = t −1+ν−µν+µk u(t), then x ∈ Cµ ((0, ∞), X ).
2262 MIAN ZHOU, YONG LIANG AND YONG ZHOU

Define an operator F by
(Fu)(t) = (F1 u)(t) + (F2 u)(t),

where
t 1−ν+µν−µk (S1 y)(t) for t ∈ (0, ∞),

(F1 u)(t) =
0 for t = 0,
t 1−ν+µν−µk (S2 y)(t) for t ∈ (0, ∞),

(F2 u)(t) =
0 for t = 0.

From (H3), there exists a constant r > 0 such that

 Z t 
µk−1
(4) sup L 2 |x0 |h(t) + L 1 t 1−ν+µν−µk
h(t) (t − s) m(s) ds ≤ r.
t∈[0,∞) 0

Let

(5)  = {u ∈ E : ∥u∥ ≤ r }, 
e = {x ∈ Cµ ((0, ∞), X ) : ∥x∥µ ≤ r }.

Then,  is a nonempty, convex and closed subset of E, and 

e is a nonempty, convex and closed subset
of Cµ ((0, ∞), X ).
Let
V := v : v(t) = h(t)(Fu)(t), u ∈  .


To prove the results in this paper we need the following lemmas.

Lemma 17. Suppose that (H1), (H2) and (H3) hold. Then, the set V is equicontinuous.

Proof. Step I. We first prove that v : v(t) = h(t)(F1 u)(t), u ∈  is equicontinuous.



Since
Z t
t 1−ν+µν−µk
t 1−ν+µν−µk Pν,µ (t)x0 = (t − s)ν(1−µ)−1 s µ−1 K µ (s)x0 ds
0(ν(1 − µ)) 0
Z 1
= (1 − z)ν(1−µ)−1 z µ−1 t µ(1−k) K µ (t z)x0 dz,
0

and noting that limt→0+ t µ(1−k) K µ (t z)x0 = 0 and (1 − z)ν(1−µ)−1 z µ−1 dz exists, we have
R1
0

lim t 1−ν+µν−µk Pν,µ (t)x0 = 0.

t→0+

Hence, for t1 = 0, t2 ∈ (0, ∞), we obtain

1−ν+µν−µk
(6) h(t2 )(F1 u)(t2 ) − h 1 (0)(F1 u)(0) ≤ h(t2 )t2 Pν,µ (t2 )x0 − 0 → 0, as t2 → 0.
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2263

For any t1 , t2 ∈ (0, ∞) and t1 < t2 , we have

(7) |h(t2 )(F1 u)(t2 ) − h(t1 )(F1 u)(t1 )|

1−ν+µν−µk 1−ν+µν−µk
= h(t2 )t2 Pν,µ (t2 )x0 − h(t1 )t1 Pν,µ (t1 )x0
1−ν+µν−µk 1−ν+µν−µk
≤ h(t2 )t2 Pν,µ (t2 )x0 − h(t1 )t2 Pν,µ (t2 )x0
1−ν+µν−µk 1−ν+µν−µk
+ h(t1 )t2 Pν,µ (t2 )x0 − h(t1 )t1 Pν,µ (t1 )x0
1−ν+µν−µk
≤ t2 Pν,µ (t2 )x0 h(t2 ) − h(t1 )
1−ν+µν−µk 1−ν+µν−µk
+ t2 Pν,µ (t2 )x0 − t1 Pν,µ (t1 )x0 h(t1 )
1−ν+µν−µk
≤ t2 Pν,µ (t2 )x0 h(t2 ) − h(t1 )
1−ν+µν−µk
+ t2 Pν,µ (t2 )x0 − Pν,µ (t1 )x0 h(t1 )
1−ν+µν−µk 1−ν+µν−µk
+ t2 − t1 Pν,µ (t1 )x0 h(t1 )
→ 0, as t2 → t1 .

Hence v : v(t) = h(t)(F1 u)(t), u ∈  is equicontinuous.



Step II. We prove that v : v(t) = h(t)(F2 u)(t), u ∈  is equicontinuous.



Let x(t) = t −1+ν−µν+µk u(t) for any u ∈ , t ∈ (0, ∞). Then x ∈ 

e.
By (H3), for ε > 0, there exists t > 0 such that
∗

t
ε
Z
(8) L 1 h(t)t 1−ν+µν−µk (t − s)µk−1 m(s) ds < for t > t ∗ .
0 2

For t1 , t2 > t ∗ , by virtue of (8), we get

(9) h(t2 )(F2 u)(t2 ) − h(t1 )(F2 u)(t1 )

Z t2 Z t1
1−ν+µν−µk 1−ν+µν−µk
≤ h(t2 )t2 Q µ (t2 − s) f (s, x(s)) ds + h(t1 )t1 Q µ (t1 − s) f (s, x(s)) ds
0 0
Z t2 Z t1
µk−1
(t1 − s)µk−1 m(s) ds < ε.
1−ν+µν−µk 1−ν+µν−µk
≤ L 1 h(t2 )t2 (t2 − s) m(s) ds + L 1 h(t1 )t1
0 0

For t1 = 0, 0 < t2 < t ∗ , by (H3), we have

Z t2
1−ν+µν−µk
(10) h(t2 )(F2 u)(t2 ) − h(0)(F2 u)(0) = h(t2 )t2 Q µ (t2 − s) f (s, x(s)) ds
0
Z t2
(t2 − s)µk−1 m(s) ds
1−ν+µν−µk
≤ L 1 h(t2 )t2
0
→ 0, as t2 → 0.
2264 MIAN ZHOU, YONG LIANG AND YONG ZHOU

For 0 < t1 < t2 ≤ t ∗ , we get

(11) h(t2 )(F2 u)(t2 ) − h(t1 )(F2 u)(t1 )
Z t2
(t2 − s)µ−1 K µ (t2 − s) f (s, x(s)) ds
1−ν+µν−µk
≤ h(t1 )t1
t1
Z t1
(t2 − s)µ−1 − (t1 − s)µ−1 K µ (t2 − s) f (s, x(s)) ds
1−ν+µν−µk
h(t1 )t1


+
0
Z t1
(t1 − s)µ−1 K µ (t2 − s) − K µ (t1 − s) f (s, x(s)) ds
1−ν+µν−µk
+ h(t1 )t1


0
Z t2
(t2 − s)µ−1 K µ (t2 − s) f (s, x(s)) ds
1−ν+µν−µk 1−ν+µν−µk
+ h(t2 )t2 − h(t1 )t1
0
≤ I1 + I2 + I3 + I4 ,
where
Z t2 Z t1
(t2 − s)µk−1 m(s) ds − (t1 − s)µk−1 m(s) ds ,
1−ν+µν−µk
I1 = L 1 h(t1 )t1
0 0
Z t1
(t1 − s)µ−1 − (t2 − s)µ−1 (t2 − s)µ(k−1) m(s) ds,
1−ν+µν−µk
I2 = 2L 1 h(t1 )t1


0
Z t1
(t1 − s)µ−1 K µ (t2 − s) − K µ (t1 − s) f (s, x(s))ds ,
1−ν+µν−µk
h(t1 )t1


I3 =
0
Z t2
(t2 − s)µk−1 m(s) ds .
1−ν+µν−µk 1−ν+µν−µk
I4 = L 1 h(t2 )t2 − h(t1 )t1
0
µk
One can deduce that limt2 →t1 I1 = 0, since I0+ m ∈ C((0, ∞), (0, ∞)). Noting that

((t1 − s)µ−1 − (t2 − s)µ−1 )(t2 − s)µ(k−1) m(s) ≤ (t1 − s)µk−1 m(s) for s ∈ [0, t1 ),
the Lebesgue dominated convergence theorem implies
Z t1
(t1 − s)µ−1 − (t2 − s)µ−1 (t2 − s)µ(k−1) m(s) ds → 0, as t2 → t1 ,


0
which implies I2 → 0 as t2 → t1 .
By (H2), for ε > 0, we have
Z t1 −ε
(t1 − s)µ−1 K µ (t2 − s) − K µ (t1 − s) L (X ) | f (s, x(s))| ds
1−ν+µν−µk
I3 ≤ h(t1 )t1
0 Z t1
(t1 − s)µ−1 K µ (t2 − s) − K µ (t1 − s) L (X ) | f (s, x(s))| ds
1−ν+µν−µk
+ h(t1 )t1
t1 −ε
Z t1
(t1 − s)µ−1 m(s) ds
1−ν+µν−µk
≤ h(t1 )t1 sup K µ (t2 − s) − K µ (t1 − s) L (X )
0 s∈[0,t1 −ε] Z t1
(t1 − s)µk−1 m(s) ds
1−ν+µν−µk
+ 2L 1 h(t1 )t1
t1 −ε
≤ I31 + I32 + I33 ,
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2265

where
Z t1
(t1 − s)µ−1 m(s) ds
1−ν+µν−µk
I31 = h(t1 )t1 sup ∥K µ (t2 − s) − K µ (t1 − s)∥L (X ) ,
0 s∈[0,t1 −ε]
Z t1 Z t1 −ε
µk−1
(t1 − ε − s)µk−1 m(s) ds ,
1−ν+µν−µk
I32 = 2L 1 Lh(t1 )t1 (t1 − s) m(s) ds −
0 0
Z t1 −ε
((t1 − ε − s)µk−1 − (t1 − s)µk−1 )m(s) ds.
1−ν+µν−µk
I33 = 2L 1 Lh(t1 )t1
0

Lemma 14 implies that I31 → 0 as t2 → t1 . Similar to the proof that I1 , I2 tend to zero, we obtain
I32 → 0 and I33 → 0 as ε → 0. Hence, I3 tends to zero as t2 → t1 . Clearly, I4 → 0 as t2 → t1 .
For 0 < t1 < t ∗ < t2 , if t2 → t1 , then t2 → t ∗ and t1 → t ∗ . Thus, for u ∈ 

(12) |h(t2 )(F2 u)(t2 ) − h(t1 )(F2 u)(t1 )|

≤ |h(t2 )(F2 u)(t2 ) − h(t ∗ )(F2 u)(t ∗ )| + |h(t ∗ )(F2 u)(t ∗ ) − h(t1 )(F2 u)(t1 )| → 0, as t2 → t1 .
Hence, for 0 ≤ t1 < t2 < ∞,

|h(t2 )(F2 u)(t2 ) − h(t1 )(F2 u)(t1 )| → 0, as t2 → t1 .

Therefore, v : v(t) = h(t)(F2 u)(t), u ∈  is equicontinuous. Furthermore, V is equicontinuous.


□
Lemma 18. Assume that (H1), (H2) and (H3) hold. Then, limt→∞ h(t)|(Fu)(t)| = 0 uniformly for u ∈ .
Proof. In fact, for any u ∈ , by (H2) and Lemma 13, we get
Z t
|(Fu)(t)| ≤ t 1−ν+µν−µk
Pν,µ (t)x0 + t 1−ν+µν−µk
Q µ (t − s) f (s, x(s)) ds
0
(13) Z t
≤ L 2 |x0 | + L 1 t 1−ν+µν−µk
(t − s)µk−1 m(s) ds.
0

Thus, we get
Z t
(14) h(t)|(Fu)(t)| ≤ L 2 h(t)|x0 | + L 1 h(t)t 1−ν+µν−µk
(t − s)µk−1 m(s) ds.
0

Consequently,

(15) h(t)|(Fu)(t)| → 0, as t → ∞,

which implies that limt→∞ h(t)|(Fu)(t)| = 0 uniformly for u ∈ . The proof is completed. □
Lemma 19. Assume that (H1), (H2) and (H3) hold. Then F ⊂ .
Proof. From Lemmas 17 and 18, we know that F ⊂ E. For any t > 0, u ∈ , from (4) and (14), we have
Z t
(16) h(t)|(Fu)(t)| ≤ L 2 |x0 |h(t) + L 1 h(t)t 1−ν+µν−µk
(t − s)µk−1 m(s) ds ≤ r.
0

For t = 0, |(Fu)(0)| = 0 < r . Therefore, F ⊂ . □

2266 MIAN ZHOU, YONG LIANG AND YONG ZHOU

Lemma 20. Suppose that (H1), (H2) and (H3) hold. Then F is continuous.

n=1 be a sequence in  which is convergent to u ∈ . Consequently,

Proof. Indeed, let {u n }∞

lim u n (t) = u(t), and lim t −1+ν−µν+µk u n (t) = t −1+ν−µν+µk u(t) for t ∈ (0, ∞).
n→∞ n→∞

Let x(t) = t −1+ν−µν+µk u(t), xn (t) = t −1+ν−µν+µk u n (t) t ∈ (0, ∞). Then x, xn ∈ 
e. In view of (H1),
we have

lim f (t, xn (t)) = lim f (t, t −1+ν−µν+µk u n (t)) = f (t, t −1+ν−µν+µk u(t)) = f (t, x(t)).
n→∞ n→∞

On the one hand, using (H3), we get for each t ∈ (0, ∞),

(t − s)µk−1 | f (s, xn (s)) − f (s, x(s))| ≤ 2(t − s)µk−1 m(s), a.e. in [0, t).

On the other hand, the function s → 2(t − s)µk−1 m(s) is integrable for s ∈ [0, t), t ∈ [0, ∞). By the
Lebesgue dominated convergence theorem, we obtain
Z t
(t − s)µk−1 | f (s, xn (s)) − f (s, x(s))| ds → 0, as n → ∞.
0

Thus, for t ∈ [0, ∞),

|h(t)(Fu n )(t) − h(t)(Fu)(t)|

Z t
≤ h(t)t |Q µ (t − s)( f (s, xn (s)) − f (s, x(s)))| ds
1−ν+µν−µk
0
Z t
≤ L 1 h(t)t 1−ν+µν−µk
(t − s)µk−1 | f (s, xn (s)) − f (s, x(s))| ds → 0, as n → ∞.
0

So, ∥Fu n − Fu∥ → 0 as n → ∞. Hence, F is continuous. The proof is completed. □

4. Main results

Theorem 21. Assume that T (t)(t > 0) is compact. Furthermore, suppose that (H1), (H2) and (H3) hold.
Then the Cauchy problem (1) has at least one mild solution.

Proof. Clearly, the problem (1) exists a mild solution x ∈ 

e if and only if the operator F has a fixed
point u ∈ , where u(t) = t 1−ν+µν−µk x(t). Hence, we only need to prove that the operator F has a
fixed point in . From Lemmas 19 and 20, we know that F ⊂  and F is continuous. In order to
prove that F is a completely continuous operator, we need to prove that F is a relatively compact
set. In view of Lemmas 17 and 18, the set V = v : v(t) = h(t)(Fu)(t), u ∈  is equicontinuous
 |h(t)(Fu)(t)| = 0 uniformly for u ∈ . According to Lemma 15, we only need to prove
and limt→∞
V (t) = v(t) : v(t) = h(t)(Fu)(t), u ∈  is relatively compact in X for t ∈ [0, ∞). Clearly, V (0) is
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2267

relatively compact in X . We only consider the case t > 0. For any ε ∈ (0, t) and δ > 0, define Fε,δ on 
as follows:
(Fε,δ u)(t) := t 1−ν+µν−µk (Sε,δ x)(t)
 Z t−εZ ∞ 
:= t 1−ν+µν−µk Pν,µ (t)x0 + µθ (t − s)µ−1 Wµ (θ )T ((t − s)µ θ ) f (s, x(s)) dθ ds .
0 δ

Thus,
h(t)(Fε,δ u)(t)
 Z t−εZ ∞ 
= h(t)t 1−ν+µν−µk Pν,µ (t)x0 +T (ε µ δ) µθ (t−s)µ−1 Wµ (θ )T ((t−s)µ θ −εµ δ) f (s, x(s)) dθ ds .
0 δ

By Theorem 3 of [8], we know that Pν,µ (t) is compact because T (t) is compact for t > 0. Further,
T (ε µ δ) is compact, so the set {h(t)(Fε,δ u)(t), u ∈ } is relatively compact in X for any ε ∈ (0, t) and
for any δ > 0. Moreover, for every u ∈ , we find
|h(t)(Fu)(t) − h(t)(Fε,δ u)(t)|
Z tZ δ
≤ h(t)t 1−ν+µν−µk
µθ (t − s)µ−1 Wµ (θ )T ((t − s)µ θ ) f (s, x(s)) dθ ds
0 0
Z t Z ∞
+ h(t)t 1−ν+µν−µk µθ (t − s)µ−1 Wµ (θ )T ((t − s)µ θ ) f (s, x(s)) dθ ds
t−ε δ
Z t Z δ
µk−1
≤ µC0 h(t)t 1−ν+µν−µk
(t − s) | f (s, x(s))| ds θ k Wµ (θ ) dθ
0 Z 0
t Z ∞
µk−1
+ µC0 h(t)t 1−ν+µν−µk
(t − s) | f (s, x(s))| ds θ k Wµ (θ ) dθ
t−ε 0
Z t Z δ
≤ µC0 h(t)t 1−ν+µν−µk (t − s)µk−1 m(s) ds θ k Wµ (θ ) dθ
0
Z t 0 Z ∞
µk−1
+ µC0 h(t)t 1−ν+µν−µk
(t − s) m(s) ds θ k Wµ (θ ) dθ
t−ε 0

→ 0, as ε → 0, δ → 0.
Thus, V (t) is also a relatively compact set in X for t ∈ [0, ∞). Therefore, Schauder fixed point theorem
implies that F has at least a fixed point u ∗ ∈ . Let x ∗ (t) = t −1+ν−µν+µk u ∗ (t). Thus
Z t
x ∗ (t) = Pν,µ (t)x0 + Q µ (t − s) f (s, x ∗ (s)) ds, t ∈ (0, ∞),
0

which implies that x∗ is a mild solution of (1). The proof is completed. □

Theorem 22. Assume that (H1), (H2) and (H3) hold. Furthermore suppose that
µk
lim I m(t) = 0.
t→∞ 0+

Then all mild solutions of the Cauchy problem (1) are attractive.
2268 MIAN ZHOU, YONG LIANG AND YONG ZHOU

Proof. Let x(t) is a mild solution of (1). From (H2) and (H3), we get
Z t
|x(t)| = Pν,µ (t)x0 + Q µ (t − s) f (s, x(s)) ds
0
µk
≤ L 2t −1+ν−µν+µk
|x0 | + L 1 0(µk)I0+ m(s) → 0, as t → ∞,
which implies that x(t) is an attractive solution. The proof is completed. □
By Theorems 21 and 22, we have the following corollaries.
Corollary 23. Assume that T (t)(t > 0) is compact and (H1) and (H2) hold. Furthermore suppose that

µk t 1−ν+µν−µk µk
(H3)′ lim t 1−ν+µν−µk I0+ m(t) = 0, lim I0+ m(t) = 0.
t→0+ t→∞ 1+t

Then the Cauchy problem (1) has at least one mild solution.
Proof. Choose h(t) = 1/(1 + t). If (H3)′ holds, then condition (H3) is satisfied. □
Corollary 24. Assume that T (t)(t > 0) is compact and (H1), (H2) and (H3)′ hold. Further suppose that
µk
lim I m(t) = 0.
t→∞ 0+

Then all mild solutions of the Cauchy problem (1) are attractive.
Corollary 25. Assume that T (t)(t > 0) is compact and (H1) and (H2) hold. Furthermore suppose that
µk µk
(H3)′′ lim t 1−ν+µν−µk I0+ m(t) = 0, lim t 1−ν+µν−µk e−t I0+ m(t) = 0.
t→0+ t→∞

Then the Cauchy problem (1) has at least one mild solution.
Proof. Choose h(t) = e−t . If (H3)′′ holds, then condition (H3) is satisfied. □
Corollary 26. Assume that T (t)(t > 0) is compact and (H1), (H2) and (H3)′′ hold. Furthermore
suppose that
µk
lim I0+ m(t) = 0.
t→∞

Then all mild solutions of the Cauchy problem (1) are attractive.
In the case that T (t)(t > 0) is noncompact, we also need the following hypothesis:
(H4) There exists constants β > 0 such that h(t)t 1−ν+µν−µk ≤ β, t ∈ [0, ∞) and
α( f (t, D)) ≤ h(t)t 1−ν+µν−µk α(D), for a.e. t ∈ [0, ∞), for any bounded D ⊆ X.
Theorem 27. Assume that (H1)–(H4) hold. Then the Cauchy problem (1) has at least one mild solution.
Proof. Let u 0 (t) = t 1−ν+µν−µk Pν,µ (t)x0 for all t ∈ [0, ∞) and u n+1 = Fu n , n = 0, 1, 2, . . . . By Lemma 19,
Fu n ∈ , for u n ∈ . We next show that the set
∞
V = vn : vn (t) = h(t)(Fu n )(t), u n ∈ 1 n=0


is relatively compact.
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2269

By Lemmas 17 and 18, the set V is equicontinuous and lim  t→∞ h(t)|(Fu n )(t)| = 0 uniformly ∞ for
u n ∈ 1 . According to Lemma 15, it suffices to prove that V (t)= vn (t):vn (t)=h(t)(Fu n )(t), u n ∈1 n=0
is relatively compact in X for t ∈ [0, ∞).
Let xn (t) = t −(1−ν+µν−µk) u n (t), t ∈ (0, ∞), n = 0, 1, 2, . . . . By the condition (H3) and Lemma 6,
we have
∞

α h(t)(Fu n )(t) n=0

 Z t ∞ 
= α h(t)t 1−ν+µν−µk
Pν,µ (t)x0 + h(t)t 1−ν+µν−µk
Q µ (t − s) f (s, xn (s)) ds
0 n=0
 Z t ∞ 
=α h(t)t 1−ν+µν−µk Q µ (t − s) f (s, xn (s)) ds
0 n=0
t
2Lh(t)t 1−ν+µν−µk
Z
(t − s)µk α f (s, {s −(1−ν+µν−µk) u n (s)}∞
n=0 ) ds


≤
0(µ) 0
t
2Lh(t)t 1−ν+µν−µk
Z
(t − s)µk h(s)s 1−ν+µν−µk α {s −(1−ν+µν−µk) u n (s)}∞


≤ n=0 ds
0(µ) 0
t
2Lh(t)t 1−ν+µν−µk
Z
(t − s)µk α h(s)u n (s)
 ∞

≤ ds.
0(µ) 0
n=0

On the other hand, by the properties of measure of noncompactness, for any t ∈ [0, ∞) we have
∞
∞

α h(t)u n (t) n=0 = α h(t)u 0 (t) ∪ h(t)u n (t) n=1
  
∞

= α h(t)u n (t) n=1


= α(V (t)).
Thus
Z t
2Lβ
(17) α(V (t)) ≤ (t − s)µk α(V (s)) ds.
0(µ) 0
Therefore, from the inequality in [6, p. 188], it follows that α(V (t)) = 0. Hence, V (t) is relatively
compact. Consequently, by Lemma 15, the set V is relatively compact, i.e., there exists a convergent
n=0 . With no confusion, let limn→∞ u n = u , u ∈ 1 .
subsequence of {u n }∞ ∗ ∗

By continuity of the operator F, we thus have

u ∗ = lim u n = lim Fu n−1 = F lim u n−1 = Fu ∗ .


n→∞ n→∞ n→∞

Let x ∗ (t) = t −(1−ν+µν−µk) u ∗ (t). This means that x∗ is a mild solution of (1). The proof is completed. □
Corollary 28. Assume that (H1), (H2), (H3)′ and (H4) hold. Then the Cauchy problem (1) has at least
one mild solution.
Proof. Choose h(t) = 1/(1 + t). If (H3)′ holds, then condition (H3) is satisfied. □
Corollary 29. Assume that (H1), (H2), (H3)′ and (H4) hold. Furthermore suppose that
µk
lim I m(t) = 0.
t→∞ 0+
Then all mild solutions of the Cauchy problem (1) are attractive.
2270 MIAN ZHOU, YONG LIANG AND YONG ZHOU

Corollary 30. Assume that (H1), (H2), (H3)′′ and (H4) hold. Then the Cauchy problem (1) has at least
one mild solution.
Proof. Choose h(t) = e−t . If (H3)′′ holds, then condition (H3) is satisfied. □
Corollary 31. Assume that (H1), (H2), (H3)′′ and (H4) hold. Furthermore suppose that
µk
lim I m(t) = 0.
t→∞ 0+

Then all mild solutions of the Cauchy problem (1) are attractive.
Example 32. Consider the following Cauchy problem on infinite interval
 H D ν,µ x(t) = Ax(t) + f 1 (t, x(t)), t ∈ (0, ∞),

0+
(18)
 I (1−µ)(1−ν) x(0) = x ,
0+ 0

where f 1 (t, · ) is Lebesgue measurable with respect to t on (0, ∞), f 1 ( · , x) is continuous with respect
to x on X , and | f 1 (t, x(t))| ≤ t −η , η ∈ (µk, 1 − ν + νµ), t ∈ (0, ∞).
Let m(t) = t −η , t > 0. Then

µk 0(1 − η) t 1−ν+µν−µk µk−η

I0+ m(t) = t µk−η , lim t 1−ν+µν−µk t µk−η = 0, lim t = 0.
0(1 + µk − η) t→0+ t→∞ 1+t
That means that the condition (H3)′ is satisfied. By Corollaries 23 and 24, the problem (18) has at least a
mild solution, and all mild solutions of (18) are attractive.
Example 33. Consider the following Cauchy problem on infinite interval
 H D ν,µ x(t) = Ax(t) + f 2 (t, x(t)), t ∈ (0, ∞),

0+
(19)
(1−µ)(1−ν)
I
0+ x(0) = x0 ,
where f 2 (t, · ) is Lebesgue measurable with respect to t on (0, ∞), f 2 ( · , x) is continuous with respect
to x on X , and | f 2 (t, x(t))| ≤ M, M ∈ (0, ∞), t ∈ (0, ∞).
Let m(t) = M, t > 0. Then

µk Mt µk Mt µk t 1−ν+µν−µk Mt µk
I0+ m(t) = , lim t 1−ν+µν−µk = 0, lim = 0.
0(1 + µk) t→0+ 0(1 + µk) t→∞ 1+t 0(1 + µk)
That means that the condition (H3)′ is satisfied. By Corollary 23, the problem (18) has at least a mild
solution.

5. Conclusions

By using the generalized Ascoli–Arzelà theorem, we investigated the existence of global mild solutions
and attractive solutions for Hilfer fractional evolution equations on infinite interval. We proved existence
theorems of global mild solutions and attractive solutions for the cases that the semigroup is compact
and noncompact respectively. In particular, we do not need to assume that f (t, · ) satisfies the Lipschitz
condition. It is worth mentioning that we also develop some new techniques, for example, structuring
 INFINITE INTERVAL PROBLEMS FOR HILFER FRACTIONAL EVOLUTION EQUATIONS 2271

the space Cµ ((0, ∞), X ) which is the key in dealing with the existence of global solutions for fractional
evolution equations on infinite intervals. The method in this paper can be applied to study infinite interval
problems for fractional evolution equations with instantaneous/noninstantaneous impulses, fractional
stochastic evolution equations.

References
[1] K. Diethelm, The analysis of fractional differential equations, Lecture Notes in Mathematics 2004, Springer, 2010.
[2] K. M. Furati, M. D. Kassim, and N. e.-. Tatar, “Existence and uniqueness for a problem involving Hilfer fractional
derivative”, Comput. Math. Appl. 64:6 (2012), 1616–1626.
[3] H. Gu and J. J. Trujillo, “Existence of mild solution for evolution equation with Hilfer fractional derivative”, Appl. Math.
Comput. 257 (2015), 344–354.
[4] W. Guo, “A generalization and application of Ascoli–Arzela theorem”, J. Systems Sci. Math. Sci. 22:1 (2002), 115–122.
[5] M. Haase, “The functional calculus for sectorial operators”, pp. 19–60 in The functional calculus for sectorial operators,
Oper. Theory Adv. Appl. 169, Birkhäuser, Basel, 2006.
[6] D. Henry, Geometric theory of semilinear parabolic equations, Lecture Notes in Mathematics 840, Springer, 1981.
[7] R. Hilfer, Applications of fractional calculus in physics, World Scientific, Singapore, 2000.
[8] A. Jaiwal and D. Bahuguna, “Hilfer fractional differential equations with almost sectorial operators”, Diff. Eq. Dyn. Syst.
(2020).
[9] K. Kavitha, V. Vijayakumar, R. Udhayakumar, and K. S. Nisar, “Results on the existence of Hilfer fractional neutral
evolution equations with infinite delay via measures of noncompactness”, Math. Methods Appl. Sci. 44:2 (2021), 1438–1455.
[10] A. A. Kilbas, H. M. Srivastava, and J. J. Trujillo, Theory and applications of fractional differential equations, North-Holland
Mathematics Studies 204, Elsevier, 2006.
[11] B. M. Liu and L. S. Liu, “The decision theorems of relative compactness for two abstract function groups in an infinite
interval and their applications”, J. Systems Sci. Math. Sci. 30:7 (2010), 1008–1019.
[12] Z. B. Liu, L. S. Liu, and J. Zhao, “The criterion for relative compactness for a class of abstract function groups on an
infinite interval and its applications”, J. Systems Sci. Math. Sci. 28:3 (2008), 370–378.
[13] F. Mainardi, P. Paraddisi, and R. Gorenflo, “Probability distributions generated by fractional diffusion equations”, pp.
312–351 in Econophysics: an emerging science, edited by J. Kertesz and I. Kondor, Kluwer, Dordrecht, 2000.
[14] F. Periago and B. Straub, “A functional calculus for almost sectorial operators and applications to abstract evolution
equations”, J. Evol. Equ. 2:1 (2002), 41–68.
[15] I. Podlubny, Fractional differential equations, Mathematics in Science and Engineering 198, Academic Press, San Diego,
CA, 1999.
[16] J. V. C. Sousa, F. Jarad, and T. Abdeljawad, “Existence of mild solutions to Hilfer fractional evolution equations in Banach
space”, Ann. Funct. Anal. 12:1 (2021), art. id. 12.
[17] M. Yang and Q. Wang, “Existence of mild solutions for a class of Hilfer fractional evolution equations with nonlocal
conditions”, Fract. Calc. Appl. Anal. 20:3 (2017), 679–705.
[18] Y. Zhou, Basic theory of fractional differential equations, World Scientific, Hackensack, NJ, 2014.
[19] Y. Zhou, Fractional evolution equations and inclusions: analysis and control, Academic Press, London, 2016.
[20] Y. Zhou and J. W. He, “Well-posedness and regularity for fractional damped wave equations”, Monatsh. Math. 194:2 (2021),
425–458.
[21] Y. Zhou and J. W. He, “Infinite interval problems for fractional evolution equations with Hilfer fractional derivative”, To
appear in Math. Meth. Appl. Sci.
2272 MIAN ZHOU, YONG LIANG AND YONG ZHOU

[22] Y. Zhou and J. N. Wang, “The nonlinear Rayleigh–Stokes problem with Riemann–Liouville fractional derivative”, Math.
Methods Appl. Sci. 44:3 (2021), 2431–2438.
[23] Y. Zhou, J. W. He, B. Ahmad, and N. Huy Tuan, “Existence and regularity results of a backward problem for fractional
diffusion equations”, Math. Methods Appl. Sci. 42:18 (2019), 6775–6790.

M IAN Z HOU : mz_math@126.com

Faculty of Innovation Engineering, Macau University of Science and Technology, Macau, China
YONG L IANG : yliang@must.edu.mo
Faculty of Innovation Engineering, Macau University of Science and Technology, Macau, China
YONG Z HOU : yzhou@xtu.edu.cn
Faculty of Innovation Engineering, Macau University of Science and Technology, Macau, China

RMJ — prepared by msp for the

Rocky Mountain Mathematics Consortium