Skip to main content

Advertisement

Estimation of unknown function of nonlinear weakly singular integral inequalities of Gronwall–Bellman–Pachpatte type

  • 345 Accesses

Abstract

Some new nonlinear weakly singular integral inequalities of Gronwall–Bellman–Pachpatte type are given. The estimations of unknown functions are obtained by analysis techniques. These estimates are very significant tools in the study of differential-integral equations.

Introduction

The integral inequality provides a useful tool for investigating the existence, uniqueness, boundedness and other qualitative properties of the solutions of differential equations and integral equations and provides an explicit bound for unknown functions. Gronwall established the essential integral inequality in 1919 [1]:

If u is a continuous function defined on the interval \(D=[\alpha , \alpha +h]\) and

$$ 0\leq u(t)\leq \int _{0}^{t}\bigl[bu(s)+a\bigr]\,ds, \quad \forall t\in D, $$

where a and b are nonnegative constants. Then

$$ 0\leq u(t)\leq (ah)\exp (bh), \quad \forall t\in D. $$

Bellman proved the Gronwall–Bellman inequality [2] in 1943, which was one of the most useful inequalities in the study of differential and integral equations:

If u and f are nonnegative continuous functions on an interval \([a, b]\) satisfying

$$ u(t)\leq c+ \int _{a}^{t} f(s)u(s)\,ds, \quad t\in [a, b], $$

for some constant \(c\geq 0\), then

$$ u(t)\leq c\exp \biggl( \int _{a}^{t}f(s)\,ds \biggr), \quad t\in [a, b]. $$

In the past few years, many scholars have extended the Gronwall–Bellman inequality and applied it to many aspects. There can be found many generalizations and analogs of it in various cases from literature (see [36]). In 2007, Jiang Fangcui and Meng Fanwei [3] investigated the estimation of the unknown function of the integral inequality:

$$ u^{p}(t)\leq c+ \int _{0}^{t}f(s)u^{p}(s)\,ds+ \int _{0}^{t}h(s)u^{q} \bigl( \sigma (s) \bigr)\,ds, \quad \forall t\in [0, \infty ). $$

In recent years, many researchers have made a great contribution to studying weakly singular integral inequalities and their applications (see [7, 8]). In [7], Xu Run and Meng Fanwei studied the following weakly singular integral inequality:

$$\begin{aligned} \textstyle\begin{cases} u^{p}(t)\leq a(t)+{\int _{t_{0}}^{t}}(t-s)^{\beta -1}b(s)u^{q}(s)\,ds+ {\int _{t_{0}}^{t}}(t-s)^{\beta -1}c(s)u^{l}(s-\gamma )\,ds, \\ \quad t\in [t_{0}, T)\subset R_{+}, \\ u(t)\leq \varphi (t), \quad t\in [t_{0}-\gamma , t_{0}). \end{cases}\displaystyle \end{aligned}$$

Ma and Pečairé [8] considered the following nonlinear singular inequalities:

$$ u^{p}(t)\leq a(t)+b(t) \int _{t_{0}}^{t}\bigl(t^{\alpha }-s^{\alpha } \bigr)^{ \beta -1}s^{\gamma -1}f(s)u^{q}(s)\,ds, \quad t\geq 0. $$

In this paper, we extend certain results that were proved in [9, 10]. Abdeldaim and Yakout [9] obtained the explicit bound to the unknown function of the following integral inequalities:

$$\begin{aligned}& u(t)\leq u_{0}+ \int _{0}^{t}f(s)u(s) \biggl[u(s)+ \int _{0}^{s}g(\lambda )u(\lambda )\,d\lambda \biggr]^{r}\,ds, \quad \forall t\in [0, \infty ), \end{aligned}$$
(1)
$$\begin{aligned}& u^{n}(t)\leq u_{0}+ \int _{0}^{t}f(s)u^{n}(s)\,ds+ \int _{0}^{t}h(s)u^{l}(s)\,ds, \quad \forall t\in [0, \infty ), \end{aligned}$$
(2)
$$\begin{aligned}& \begin{aligned}[b] u(t) &\leq u_{0}+ \biggl[ \int _{0}^{t}f(s)u(s)\,ds \biggr]^{2}+ \int _{0} ^{t}f(s)u(s) \biggl[u(s) \\ &\quad{} +2 \int _{0}^{s}f(\lambda )u(\lambda )\,d\lambda \biggr]\,ds, \quad \forall t \in [0, \infty ). \end{aligned} \end{aligned}$$
(3)

In 2012, Wu-sheng Wang [10] studied the following integral inequality:

$$\begin{aligned} u(t)\leq u_{0}+ \int _{0}^{\alpha (t)}f(s)\varphi _{1} \bigl(u(s) \bigr) \biggl[u(s)+ \int _{0}^{s}g(\lambda )\varphi _{2} \bigl(u(\lambda ) \bigr)\,d \lambda \biggr]^{r}\,ds. \end{aligned}$$
(4)

In 2017, Wang and Huang [11] studied the weakly singular integral inequality:

$$\begin{aligned} \begin{aligned}[b] u(t) &\leq b(t)+ \int _{a}^{t}(t-s)^{\beta -1}f_{1}(s) \varphi _{1} \bigl(u(s) \bigr) \biggl\{ u(s)+ \int _{a}^{s}(s-\tau )^{\beta -1}f_{2}( \tau ) \\ &\quad {}\times \varphi _{2} \bigl(u(\tau ) \bigr) \biggl[u(\tau ) + \int _{a}^{ \tau }(\tau -\xi )^{\beta -1}f_{3}( \xi )\varphi _{3} \bigl(u(\xi ) \bigr)\,d \xi \biggr]\,d\tau \biggr\} \,ds. \end{aligned} \end{aligned}$$
(5)

The aim of this paper is to extend certain results that were proved in [9, 10], and generalize (1)–(4) to some weakly singular integral inequalities. The upper bound estimations of the unknown functions are given by means of discrete Jensen inequality, the Hölder integral inequality and amplification techniques. Furthermore, we apply our result to integral equations for estimation.

Preliminaries and basic lemmas

In this section,we give some lemmas.

Lemma 2.1

([12]; Hölder integral inequality)

We assume that \(f(x)\)and \(g(x)\)are nonnegative continuous functions defined on \([c, d]\)and \(p>1\), \(\frac{1}{p}+\frac{1}{q}=1\), then we have

$$\begin{aligned} \int _{c}^{d}f(x)g(x)\,dx\leq \biggl( \int _{c}^{d}f^{p}(x)\,dx \biggr)^{ \frac{1}{p}} \biggl( \int _{c}^{d}g^{q}(x)\,dx \biggr)^{\frac{1}{q}}. \end{aligned}$$
(6)

Lemma 2.2

([12])

If \(\beta \in (0, \frac{1}{2}]\), \(p=1+ \beta \), then

$$\begin{aligned} \int _{t_{0}}^{t}(t-s)^{p(\beta -1)}e^{ps} \,ds\leq \frac{e^{pt}}{p^{1+p( \beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr), \quad \forall t_{0}\in R_{+}, \end{aligned}$$
(7)

where \(\varGamma (\beta )=\int _{0}^{\infty }\tau ^{\beta -1}e^{-\tau }\,d \tau \)is the Gamma function.

Lemma 2.3

([13]; Discrete Jensen inequality)

Let \(A_{1}, A_{2},\ldots, A_{n}\), \(l>1\)be nonnegative real-valued constants andna constant natural number, then

$$\begin{aligned} (A_{1}+A_{2}+\cdots +A_{n})^{l}\leq n^{l-1}\bigl(A_{1}^{l}+A_{2}^{l}+ \cdots +A_{n}^{l}\bigr). \end{aligned}$$
(8)

Main results

In this section, we discuss some nonlinear weakly singular integral inequalities. Throughout this paper, let \(I=[t_{0}, \infty )\).

Theorem 3.1

We assume that \(\beta \in (0, \frac{1}{2}]\)is a constant, \(u(t)\), \(f(t)\)and \(g(t)\)are nonnegative and nondecreasing real-valued continuous functions defined onIand satisfy the inequality

$$\begin{aligned} \begin{gathered} \begin{aligned} u(t) &\leq u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)u(s) \biggl[u(s) \\ &\quad {}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda )u(\lambda )\,d \lambda \biggr]^{r}\,ds, \quad \forall t\in I, \end{aligned} \\ {2^{(q-1)(2r+1)}u_{0}^{rq}rP(t) \int _{t_{0}}^{t}f^{q}(s)e^{-qs}C_{1}(s) \,ds< 1, \quad \forall t\in I,} \end{gathered} \end{aligned}$$
(9)

where \(u_{0}\)andrare positive constants. Then

$$\begin{aligned} u(t)\leq 2^{\frac{q-1}{q}}u_{0} \biggl[\exp \biggl( \int _{t_{0}}^{t}2^{(q-1)(r+1)}P(s)f ^{q}(s)e^{-qs}B_{1}(s)\,ds \biggr) \biggr]^{\frac{1}{q}}, \quad \forall t\in I, \end{aligned}$$
(10)

where

$$\begin{aligned}& P(t)= \biggl[\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr) \biggr]^{\frac{q}{p}}, \end{aligned}$$
(11)
$$\begin{aligned}& {B_{1}(t)=\frac{(2^{q-1}u_{0}^{q})^{r}C_{1}(t)}{1-2^{(q-1)(2r+1)}u _{0}^{rq}rP(t)\int _{t_{0}}^{t}f^{q}(s)e^{-qs}C_{1}(s)\,ds}, \quad \forall t \in I,} \end{aligned}$$
(12)
$$\begin{aligned}& C_{1}(t)=\exp \biggl(rP(t) \int _{t_{0}}^{t}g^{q}(s)e^{-qs}\,ds \biggr), \quad \forall t\in I, \end{aligned}$$
(13)

and \(p=1+\beta \), \(q=\frac{1+\beta }{\beta }\).

Proof

Using (6) and (7), from (9) we get

$$\begin{aligned} u(t) \leq & u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}f(s)e^{-s}u(s) \biggl[u(s) \\ &{}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda )u(\lambda )\,d \lambda \biggr]^{r}\,ds \\ \leq& u_{0}+ \biggl[\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr) \biggr]^{\frac{1}{p}} \biggl\{ \int _{t_{0}}^{t}f ^{q}(s)e^{-qs}u^{q}(s) \biggl[u(s) \\ &{}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda )u(\lambda )\,d \lambda \biggr]^{rq}\,ds \biggr\} ^{\frac{1}{q}}, \quad \forall t\in I. \end{aligned}$$
(14)

Using (8), from (14) we have

$$\begin{aligned} u^{q}(t) \leq &2^{q-1}u_{0}^{q}+2^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}u^{q}(s) \biggl[u(s) \\ &{}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda )u(\lambda )\,d \lambda \biggr]^{rq}\,ds, \quad \forall t\in I, \end{aligned}$$
(15)

where \(P(t)\) is defined by (11).

Also using (6), (7) and (8), \(\forall t\in I\), we have

$$\begin{aligned}& \biggl[u(s)+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda )u(\lambda )\,d\lambda \biggr]^{rq} \\& \quad \leq \biggl[2^{q-1}u^{q}(s)+2^{q-1}P(s) \int _{t_{0}}^{s}g^{q}(\lambda )e ^{-q\lambda }u^{q}(\lambda )\,d\lambda \biggr]^{r}. \end{aligned}$$
(16)

Substituting (16) into (15) we get

$$\begin{aligned} u^{q}(t) \leq &2^{q-1}u_{0}^{q}+2^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}u^{q}(s) \biggl[2^{q-1}u^{q}(s) \\ &{}+2^{q-1}P(s) \int _{t_{0}}^{s}g^{q}(\lambda )e^{-q\lambda }u^{q}( \lambda )\,d\lambda \biggr]^{r}\,ds \\ \leq &2^{q-1}u_{0}^{q}+2^{(q-1)(r+1)}P(T) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}u^{q}(s) \biggl[u^{q}(s) \\ &{}+P(T) \int _{t_{0}}^{s}g^{q}(\lambda )e^{-q\lambda }u^{q}(\lambda )\,d \lambda \biggr]^{r}\,ds, \end{aligned}$$
(17)

for all \(t\in [t_{0}, T]\), where \(t_{0}< T<\infty \) is chosen arbitrarily.

Let \(z_{1}(t)\) equal the right hand side in (17), we have \(z_{1}(t_{0})=2^{q-1}u_{0}^{q}\) and

$$\begin{aligned} u(t)\leq z_{1}^{\frac{1}{q}}(t). \end{aligned}$$
(18)

Differentiating \(z_{1}(t)\) with respect to t, and using (18) we obtain

$$\begin{aligned} \frac{dz_{1}(t)}{dt} =&2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}u^{q}(t) \biggl[u^{q}(t) \\ &{}+P(T) \int _{t_{0}}^{t}g^{q}(\lambda )e^{-q\lambda }u^{q}(\lambda )\,d \lambda \biggr]^{r} \\ \leq &2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}z_{1}(t) \biggl[z_{1}(t) \\ &{}+P(T) \int _{t_{0}}^{t}g^{q}(\lambda )e^{-q\lambda }z_{1}(\lambda )\,d \lambda \biggr]^{r}, \quad \forall t\in [t_{0}, T]. \end{aligned}$$
(19)

Letting \(z_{2}(t)=z_{1}(t)+P(T)\int _{t_{0}}^{t}g^{q}(\lambda )e^{-q \lambda }z_{1}(\lambda )\,d\lambda \), we have \(z_{2}(t_{0})=2^{q-1}u _{0}^{q}\) and

$$\begin{aligned} z_{1}(t)\leq z_{2}(t). \end{aligned}$$
(20)

Differentiating \(z_{2}(t)\) with respect to t, and using (19) and (20) we have

$$\begin{aligned} \frac{dz_{2}(t)}{dt} =&\frac{dz_{1}(t)}{dt}+P(T)g^{q}(t)e^{-qt}z_{1}(t) \\ \leq &2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}z_{2}^{r+1}(t)+P(T)g^{q}(t)e ^{-qt}z_{2}(t), \end{aligned}$$

for all \(t\in [t_{0}, T]\), but \(z_{2}(t)>0\) (where \(u_{0}>0\)), then we have

$$\begin{aligned} z_{2}^{-(r+1)}(t)\frac{dz_{2}(t)}{dt}-P(T)g^{q}(t)e^{-qt}z_{2}^{-r}(t) \leq 2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}. \end{aligned}$$
(21)

If we let

$$\begin{aligned} S_{1}(t)=z_{2}^{-r}(t), \quad \forall t\in [t_{0}, T], \end{aligned}$$
(22)

then we get \(S_{1}(t_{0})=z_{2}^{-r}(t_{0})=(2^{q-1}u_{0}^{q})^{-r}\), thus from (21) and (22) we obtain

$$ \frac{dS_{1}(t)}{dt}+rP(T)g^{q}(t)e^{-qt}S_{1}(t) \geq -r2^{(q-1)(r+1)}P(T)f ^{q}(t)e^{-qt}, \quad \forall t\in [t_{0}, T]. $$

The above inequality implies the following estimation for \(S_{1}(t)\):

$$\begin{aligned} S_{1}(t)\geq \frac{(2^{q-1}u_{0}^{q})^{-r}-\int _{t_{0}}^{t}r2^{(q-1)(r+1)}P(T)f ^{q}(s)e^{-qs}c_{1}(s)\,ds}{c_{1}(t)}, \quad \forall t\in [t_{0}, T], \end{aligned}$$

where \(c_{1}(t)=\exp \int _{t_{0}}^{t}rP(T)g^{q}(s)e^{-qs}\,ds\).

Let \(t=T\), because \(t_{0}< T<\infty \) is chosen arbitrarily, we get

$$\begin{aligned} {S_{1}(t)\geq \frac{1-2^{(q-1)(2r+1)}u_{0}^{rq}rP(t)\int _{t_{0}}^{t}f ^{q}(s)e^{-qs}C_{1}(s)\,ds}{(2^{q-1}u_{0}^{q})^{r}C_{1}(t)}, \quad \forall t \in I,} \end{aligned}$$
(23)

where \(C_{1}(t)\) is defined by (13). Then from (23) in (22), we have

$$ {z_{2}^{r}(t)\leq \frac{(2^{q-1}u_{0}^{q})^{r}C_{1}(t)}{1-2^{(q-1)(2r+1)}u _{0}^{rq}rP(t)\int _{t_{0}}^{t}f^{q}(s)e^{-qs}C_{1}(s)\,ds}=B_{1}(t), \quad \forall t\in I,} $$

where \(B_{1}(t)\) is defined by (12), thus from (19) we have

$$ \frac{dz_{1}(t)}{dt}\leq 2^{(q-1)(r+1)}P(t)f^{q}(t)e^{-qt}z_{1}(t)B _{1}(t), \quad \forall t\in I, $$

the above inequality implies an estimation for \(z_{1}(t)\) as in the following:

$$\begin{aligned} z_{1}(t)\leq 2^{q-1}u_{0}^{q} \exp \biggl( \int _{t_{0}}^{t}2^{(q-1)(r+1)} {P(s)}f^{q}(s)e^{-qs}B_{1}(s) \,ds \biggr), \quad \forall t\in I, \end{aligned}$$
(24)

and from (18) and (24) we have

$$ u(t)\leq 2^{\frac{q-1}{q}}u_{0} \biggl[\exp \biggl( \int _{t_{0}}^{t}2^{(q-1)(r+1)} {P(s)}f^{q}(s)e^{-qs}B_{1}(s) \,ds \biggr) \biggr]^{\frac{1}{q}}, \quad \forall t\in I. $$

 □

Theorem 3.2

We assume that \(\beta \in (0, \frac{1}{2}]\)is a constant, \(u(t)\), \(f(t)\)and \(g(t)\)are nonnegative and nondecreasing real-valued continuous functions defined on I. Suppose \(\varphi _{1}, \varphi _{2}\in C^{1}(I, I)\)are increasing functions with \(\varphi _{i}(t)>0\), \(\forall t>t_{0}\), \(i=1,2\), and \(\frac{\varphi _{1}}{\varphi _{2}}\)is increasing and nonnegative function too. If the inequality

$$\begin{aligned} u(t) \leq &u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)\varphi _{1} \bigl(u(s) \bigr) \biggl[u(s) \\ &{}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda ) \varphi _{2} \bigl(u(\lambda ) \bigr)\,d\lambda \biggr]^{r} \,ds, \quad \forall t\in I, \end{aligned}$$
(25)

is satisfied, where \(u_{0}\)andrare positive constants, then

$$\begin{aligned} u(t)\leq \bigl\{ \varPhi _{1}^{-1} \bigl[\varPhi _{2}^{-1}\bigl[U(t)\bigr] \bigr] \bigr\} ^{\frac{1}{q}}, \quad \forall t\in I, \end{aligned}$$
(26)

where \(P(t)\)is defined as (11), \(p=1+\beta \), \(q=\frac{1+\beta }{ \beta }\)and

$$\begin{aligned}& \begin{aligned}[b] U(t) &=\varPhi _{2} \biggl[\varPhi _{1} \bigl(2^{q-1}u_{0}^{q}\bigr)+ \int _{t_{0}}^{t}P(s)g ^{q}(s)e^{-qs} \,ds \biggr] \\ &\quad {}+ \int _{t_{0}}^{t}2^{(q-1)(r+1)}P(s)f^{q}(s)e^{-qs} \,ds, \quad \forall t \in I, \end{aligned} \end{aligned}$$
(27)
$$\begin{aligned}& \varPhi _{1}= \int _{1}^{k}\frac{ds}{\varphi _{2}^{q}(s^{\frac{1}{q}})},\quad k>0, \end{aligned}$$
(28)
$$\begin{aligned}& \varPhi _{2}= \int _{1}^{k}\frac{\varphi _{2}^{q} [[\varPhi _{1}^{-1}(s)]^{ \frac{1}{q}} ]}{\varphi _{1}^{q} [[\varPhi _{1}^{-1}(s)]^{ \frac{1}{q}} ][\varPhi _{1}^{-1}(s)]^{r}}\,ds, \quad k>0, \end{aligned}$$
(29)

and satisfied \(U(t)\leq \varPhi _{2}(\infty )\), \(\varPhi _{2}^{-1}[U(t)]\leq \varPhi _{1}(\infty )\).

Proof

Using (6) and (7), from (25) we get

$$\begin{aligned} u(t) \leq & u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}f(s)e^{-s} \varphi _{1} \bigl(u(s) \bigr) \biggl[u(s) \\ & {} + \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda ) \varphi _{2} \bigl(u(\lambda ) \bigr)\,d\lambda \biggr]^{r} \,ds \\ \leq & u_{0}+ \biggl[\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p( \beta -1) \bigr) \biggr]^{\frac{1}{p}} \biggl\{ \int _{t_{0}}^{t}f^{q}(s)e ^{-qs} \varphi _{1}^{q} \bigl(u(s) \bigr) \biggl[u(s) \\ & {} + \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda ) \varphi _{2} \bigl(u(\lambda ) \bigr)\,d\lambda \biggr]^{rq} \,ds \biggr\} ^{\frac{1}{q}},\quad \forall t\in I. \end{aligned}$$
(30)

Using (8), from (30) we have

$$\begin{aligned} u^{q}(t) \leq &2^{q-1}u_{0}^{q}+2^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs} \varphi _{1}^{q} \bigl(u(s) \bigr) \biggl[u(s) \\ &{}+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda ) \varphi _{2} \bigl(u(\lambda ) \bigr)\,d\lambda \biggr]^{rq} \,ds,\quad \forall t\in I, \end{aligned}$$
(31)

where \(P(t)\) is defined by (11).

Also using (6), (7) and (8), \(\forall t\in I\), we have

$$\begin{aligned}& \biggl[u(s)+ \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}g(\lambda ) \varphi _{2} \bigl(u(\lambda ) \bigr)\,d\lambda \biggr]^{rq} \\& \quad \leq \biggl[2^{q-1}u^{q}(s)+2^{q-1}P(s) \int _{t_{0}}^{s}g^{q}(\lambda )e ^{-q\lambda }\varphi _{2}^{q} \bigl(u(\lambda ) \bigr)\,d \lambda \biggr]^{r}. \end{aligned}$$
(32)

Substituting (32) into (31) we get

$$\begin{aligned} u^{q}(t) \leq &2^{q-1}u_{0}^{q}+ {2^{q-1}}P(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs} \varphi _{1}^{q} \bigl(u(s) \bigr) \biggl[2^{q-1}u^{q}(s) \\ &{}+2^{q-1}P(s) \int _{t_{0}}^{s}g^{q}(\lambda )e^{-q\lambda }\varphi _{2}^{q} \bigl(u(\lambda ) \bigr) \,d\lambda \biggr]^{r}\,ds \\ \leq &2^{q-1}u_{0}^{q}+2^{(q-1)(r+1)}P(T) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs} \varphi _{1}^{q} \bigl(u(s) \bigr) \biggl[u^{q}(s) \\ &{}+P(T) \int _{t_{0}}^{s}g^{q}(\lambda )e^{-q\lambda }\varphi _{2}^{q} \bigl(u(\lambda ) \bigr) \,d\lambda \biggr]^{r}\,ds, \end{aligned}$$
(33)

for all \(t\in [t_{0}, T]\), where \(t_{0}< T<\infty \) is chosen arbitrarily.

Let \(z_{3}(t)\) equal the right hand side in (33), we have \(z_{3}(t_{0})=2^{q-1}u_{0}^{q}\) and

$$ u(t)\leq z_{3}^{\frac{1}{q}}(t). $$
(34)

Differentiating \(z_{3}(t)\) with respect to t, and using (34) we obtain

$$\begin{aligned} \frac{dz_{3}(t)}{dt} =&2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt} \varphi _{1} ^{q} \bigl(u(t) \bigr) \biggl[u^{q}(t) \\ &{}+P(T) \int _{t_{0}}^{t}g^{q}(\lambda )e^{-q\lambda }\varphi _{2}^{q} \bigl(u(\lambda ) \bigr) \,d\lambda \biggr]^{r} \\ \leq &2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}\varphi _{1}^{q} \bigl(z_{3}^{ \frac{1}{q}}(t) \bigr) \biggl[z_{3}(t) \\ &{}+P(T) \int _{t_{0}}^{t}g^{q}(\lambda )e^{-q\lambda }\varphi _{2}^{q} \bigl(z_{3}^{\frac{1}{q}}( \lambda ) \bigr)\,d\lambda \biggr]^{r},\quad \forall t\in [t_{0}, T]. \end{aligned}$$
(35)

Letting \(z_{4}(t)=z_{3}(t)+P(T)\int _{t_{0}}^{t}g^{q}(\lambda )e^{-q \lambda }\varphi _{2}^{q} (z_{3}^{\frac{1}{q}}(\lambda ) )\,d \lambda \), then we have \(z_{4}(t_{0})=2^{q-1}u_{0}^{q}\) and

$$ z_{3}(t)\leq z_{4}(t). $$
(36)

Differentiating \(z_{4}(t)\) with respect to t, and using (35) and (36) we have

$$\begin{aligned} \frac{dz_{4}(t)}{dt} =&\frac{dz_{3}(t)}{dt}+P(T)g^{q}(t)e^{-qt} \varphi _{2}^{q} \bigl(z_{4}^{\frac{1}{q}}(t) \bigr) \\ \leq &2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}\varphi _{1}^{q} \bigl(z_{4}^{ \frac{1}{q}}(t) \bigr)z_{4}^{r}(t) \\ &{}+P(T)g^{q}(t)e^{-qt}\varphi _{2}^{q} \bigl(z_{4}^{\frac{1}{q}}(t) \bigr), \quad \forall t\in [t_{0}, T]. \end{aligned}$$
(37)

Since \(\varphi _{2}^{q} (z_{4}^{\frac{1}{q}}(t) )>0\), \(\forall t>t _{0}\), we have

$$\begin{aligned} \frac{\frac{dz_{4}(t)}{dt}}{\varphi _{2}^{q} (z_{4}^{\frac{1}{q}}(t) )} \leq & P(T)g^{q}(t)e^{-qt} \\ &{}+\frac{2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}\varphi _{1}^{q} (z_{4} ^{\frac{1}{q}}(t) )z_{4}^{r}(t)}{\varphi _{2}^{q} (z_{4}^{ \frac{1}{q}}(t) )}, \quad \forall t\in [t_{0}, T]. \end{aligned}$$
(38)

By taking \(t=s\) in (38) and integrating it from \(t_{0}\) to t, and using (28) we get

$$\begin{aligned} \varPhi _{1} \bigl(z_{4}(t) \bigr) \leq & \varPhi _{1} \bigl(z_{4}(t_{0}) \bigr)+ \int _{t_{0}}^{t}P(T)g^{q}(s)e^{-qs} \,ds \\ &{}+ \int _{t_{0}}^{t}\frac{2^{(q-1)(r+1)}P(T)f^{q}(s)e^{-qs}\varphi _{1} ^{q} (z_{4}^{\frac{1}{q}}(s) )z_{4}^{r}(s)}{\varphi _{2}^{q} (z_{4}^{\frac{1}{q}}(s) )}\,ds \\ \leq &\varPhi _{1} \bigl(z_{4}(t_{0}) \bigr)+ \int _{t_{0}}^{T}P(T)g^{q}(s)e ^{-qs}\,ds \\ &{}+ \int _{t_{0}}^{t}\frac{2^{(q-1)(r+1)}P(T)f^{q}(s)e^{-qs}\varphi _{1} ^{q} (z_{4}^{\frac{1}{q}}(s) )z_{4}^{r}(s)}{\varphi _{2}^{q} (z_{4}^{\frac{1}{q}}(s) )}\,ds \end{aligned}$$
(39)

for all \(t\in [t_{0}, T]\), where \(\varPhi _{1}\) is defined by (28).

Let \(z_{5}(t)\) equal the right hand side in (39), we have \(z_{5}(t_{0})=\varPhi _{1}(2^{q-1}u_{0}^{q})+\int _{t_{0}}^{T}P(T)g^{q}(s) e ^{-qs}\,ds\), and

$$ z_{4}(t)\leq \varPhi _{1}^{-1} \bigl(z_{5}(t) \bigr). $$
(40)

Differentiating \(z_{5}(t)\) with respect to t, and using (40) we get

$$\begin{aligned} \frac{dz_{5}(t)}{dt} = &\frac{2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}\varphi _{1} ^{q} (z_{4}^{\frac{1}{q}}(t) )z_{4}^{r}(t)}{\varphi _{2}^{q} (z_{4}^{\frac{1}{q}}(t) )} \\ \leq &\frac{2^{(q-1)(r+1)}P(T)f^{q}(t)e^{-qt}\varphi _{1} ^{q} [ (\varPhi _{1}^{-1} (z_{5}(t) ) )^{\frac{1}{q}} ] [\varPhi _{1}^{-1} (z_{5}(t) ) ]^{r}}{\varphi _{2}^{q} [ (\varPhi _{1}^{-1} (z_{5}(t) ) )^{\frac{1}{q}} ]}, \end{aligned}$$
(41)

for all \(t\in [t_{0}, T]\).

By taking \(t=s\) in (41) and integrating it from \(t_{0}\) to t, and using (29) we get

$$ \varPhi _{2} \bigl(z_{5}(t) \bigr)\leq \varPhi _{2} \bigl(z_{5}(t_{0}) \bigr)+ \int _{t_{0}}^{t}2^{(q-1)(r+1)}P(T)f^{q}(s)e^{-qs} \,ds, \quad \forall t\in [t_{0}, T]. $$
(42)

Let \(t=T\), from (42) we have

$$ \varPhi _{2} \bigl(z_{5}(T) \bigr)\leq \varPhi _{2} \bigl(z_{5}(t_{0}) \bigr)+ \int _{t_{0}}^{T}2^{(q-1)(r+1)}P(T)f^{q}(s)e^{-qs} \,ds, \quad \forall t\in [t_{0}, T]. $$
(43)

Because \(t_{0}< T<\infty \) is chosen arbitrarily, from (34), (36) and (40), we have

$$ u(t)\leq \bigl\{ \varPhi _{1}^{-1} \bigl[\varPhi _{2}^{-1}\bigl[U(t)\bigr] \bigr] \bigr\} ^{\frac{1}{q}}, \quad \forall t\in I, $$

where \(U(t)\) is defined by (27). □

Remark

It is interesting to note that in the special case when \(\varphi _{1}(t)=t\) and \(\varphi _{2}(t)=t\) the inequality given in Theorem 3.2 reduces to Theorem 3.1.

Theorem 3.3

We assume that \(\beta \in (0, \frac{1}{2}]\)is a constant, \(u(t)\), \(f(t)\)and \(h(t)\)are nonnegative real-valued continuous functions defined onI, and they satisfy the inequality

$$\begin{aligned} u^{n}(t)\leq u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)u^{n}(s) \,ds+ \int _{t_{0}}^{t}(t-s)^{\beta -1}h(s)u^{l}(s) \,ds, \quad \forall t\in I, \end{aligned}$$
(44)

where \(u_{0}>0\), and \(n>l\geq 0\)are constants. Then

$$\begin{aligned} u(t) \leq &\exp \biggl(\frac{1}{nq}\omega _{1}(t) \biggr) \biggl\{ \bigl(3^{q-1}u _{0}^{q} \bigr)^{n_{1}}+3^{q-1}P(t)n_{1} \\ &{}\times \int _{t_{0}}^{t}h^{q}(s)e^{-qs} \bigl[\exp \bigl(-n_{1}\omega _{1}(s) \bigr) \bigr] \biggr\} ^{\frac{1}{(n-l)q}}, \quad \forall t\in I, \end{aligned}$$
(45)

where \(P(t)\)is defined by (11), \(n_{1}=\frac{n-l}{n}\), \(p=1+ \beta \), \(q=\frac{1+\beta }{\beta }\)and

$$ \omega _{1}(t)=3^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e^{-qs}\,ds, \quad \forall t \in I. $$
(46)

Proof

Using (6) and (7), from (44) we get

$$\begin{aligned} u^{n}(t) \leq & u_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}f(s)e^{-s}u ^{n}(s)\,ds \\ & {} + \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}h(s)e^{-s}u^{l}(s) \,ds \\ \leq & u_{0}+ \biggl[\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr) \biggr]^{\frac{1}{p}} \biggl[ \int _{t_{0}}^{t}f ^{q}(s)e^{-qs}u^{nq}(s) \,ds \biggr]^{\frac{1}{q}} \\ & {} + \biggl[\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr) \biggr]^{\frac{1}{p}} \biggl[ \int _{t_{0}}^{t}h^{q}(s)e^{-qs}u^{lq}(s) \,ds \biggr]^{\frac{1}{q}}, \\ &\forall t\in I. \end{aligned}$$
(47)

Using (8) and from (47) we have

$$\begin{aligned} u^{nq}(t) \leq &3^{q-1}u_{0}^{q}+3^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}u^{nq}(s)\,ds \\ & {} +3^{q-1}P(t) \int _{t_{0}}^{t}h^{q}(s)e^{-qs}u^{lq}(s) \,ds \\ \leq &3^{q-1}u_{0}^{q}+3^{q-1}P(T) \biggl[ \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}u^{nq}(s)\,ds \\ & {} + \int _{t_{0}}^{t}h^{q}(s)e^{-qs}u^{lq}(s) \,ds \biggr], \end{aligned}$$
(48)

for all \(t\in [t_{0}, T]\), where \(t_{0}< T<\infty \) is chosen arbitrarily, and \(P(t)\) is defined by (11).

Let \(z_{6}^{n}(t)\) equal the right hand side in (48), we have \(z_{6}(t_{0})=(3^{q-1}u_{0}^{q})^{\frac{1}{n}}\) and

$$ u(t)\leq z_{6}^{\frac{1}{q}}(t). $$
(49)

Differentiating \(z_{6}^{n}(t)\) with respect to t, and using (48) and (49) we have

$$\begin{aligned} nz_{6}^{n-1}(t)\frac{dz_{6}(t)}{dt} = &3^{q-1}P(T) \bigl[f^{q}(t)e^{-qt}u^{nq}(t)+h^{q}(t)e ^{-qt}u^{lq}(t) \bigr] \\ \leq &3^{q-1}P(T) \bigl[f^{q}(t)e^{-qt}z_{6}^{n}(t)+h^{q}(t)e ^{-qt}z_{6}^{l}(t) \bigr], \quad \forall t\in [t_{0}, T]. \end{aligned}$$
(50)

Since \(u_{0}>0\) we have \(z_{6}(t)>0\). Thus, we have

$$\begin{aligned} nz_{6}^{n-l-1}(t)\frac{dz_{6}(t)}{dt} \leq &3^{q-1}P(T)f^{q}(t)e^{-qt}z _{6}^{n-l}(t) \\ & {} +3^{q-1}P(T)h^{q}(t)e^{-qt}, \quad \forall t \in [t_{0}, T], \end{aligned}$$
(51)

if we let

$$\begin{aligned} z_{6}^{n-l}(t)=z_{7}(t), \quad \forall t\in [t_{0}, T], \end{aligned}$$
(52)

then we have \(z_{7}(t_{0})=(3^{q-1}u_{0}^{q})^{n_{1}}\), and \(nz_{6}^{n-l-1}(t)\frac{dz_{6}(t)}{dt}=\frac{1}{n_{1}} \frac{dz_{7}(t)}{dt}\), thus from (51) we obtain

$$\begin{aligned} \frac{dz_{7}(t)}{dt}\leq 3^{q-1}P(T)n_{1}f^{q}(t)e^{-qt}z_{7}(t)+3^{q-1}P(T)n _{1}h^{q}(t)e^{-qt}, \quad \forall t\in [t_{0}, T]. \end{aligned}$$
(53)

The inequality (53) implies the estimation for \(z_{7}(t)\), as

$$\begin{aligned} z_{7}(t) \leq &\exp \biggl(n_{1}3^{q-1}P(T) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}\,ds \biggr) \biggl\{ \bigl(3^{q-1}u_{0}^{q} \bigr)^{n_{1}}+3^{q-1}P(T)n_{1} \\ & {} \times \int _{t_{0}}^{t}h^{q}(s)e^{-qs} \biggl[\exp \biggl(-n_{1}3^{q-1}P(T) \int _{t_{0}}^{s}f^{q}(\lambda )e^{-q\lambda }\,d\lambda \biggr) \biggr]\,ds \biggr\} , \end{aligned}$$
(54)

for all \(t\in [t_{0}, T]\).

Letting \(t=T\), because \(t_{0}< T<\infty \) is chosen arbitrarily, then we get

$$\begin{aligned} z_{7}(t) \leq &\exp \bigl(n_{1}\omega _{1}(t) \bigr) \biggl\{ \bigl(3^{q-1}u _{0}^{q} \bigr)^{n_{1}}+3^{q-1}P(t)n_{1} \\ & {} \times \int _{t_{0}}^{t}h^{q}(s)e^{-qs} \bigl[\exp \bigl(-n_{1}\omega _{1}(s) \bigr) \bigr]\,ds \biggr\} , \quad \forall t\in I, \end{aligned}$$
(55)

where \(\omega _{1}(t)\) is defined by (46).

Then from (55) in (52), we have

$$\begin{aligned} z_{6}(t) \leq &\exp \biggl(\frac{1}{n}\omega _{1}(t) \biggr) \biggl\{ \bigl(3^{q-1}u _{0}^{q} \bigr)^{n_{1}}+3^{q-1}P(t)n_{1} \\ & {} \times \int _{t_{0}}^{t}h^{q}(s)e^{-qs} \bigl[\exp \bigl(-n_{1}\omega _{1}(s) \bigr) \bigr]\,ds \biggr\} ^{\frac{1}{n-l}}, \quad \forall t\in I, \end{aligned}$$
(56)

thus from (49) we can get (45). □

Theorem 3.4

We assume that \(\beta \in (0, \frac{1}{2}]\)is a constant, \(u(t)\)and \(f(t)\)are nonnegative real-valued continuous functions defined onI, and they satisfy the inequality

$$\begin{aligned} \begin{gathered} \begin{aligned} u(t) &\leq u_{0}+ \biggl[ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)u(s)\,ds \biggr]^{2}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)u(s) \biggl[u(s) \\ &\quad {} +2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )u(\lambda )\,d \lambda \biggr]\,ds, \quad \forall t\in I, \end{aligned} \\ 108^{q-1}P^{2}(t)u_{0}^{q} \int _{t_{0}}^{t}f^{q}(s)e^{-qs} \bigl[\exp \bigl(\omega _{2}(s) \bigr) \bigr]\,ds< 1, \end{gathered} \end{aligned}$$
(57)

where \(u_{0}>0\)is constant. Then

$$\begin{aligned} u(t)\leq 3^{\frac{q-1}{q}}u_{0}\exp \biggl[ \frac{1}{q} \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}B_{2}(s) \,ds \biggr], \quad \forall t\in I, \end{aligned}$$
(58)

where

$$\begin{aligned} B_{2}(t)=\frac{18^{q-1}P(t)u_{0}^{q}\exp (\omega _{2}(t) )}{1-108^{q-1}P ^{2}(t)u_{0}^{q}\int _{t_{0}}^{t}f^{q}(s)e^{-qs} [\exp (\omega _{2}(s) ) ]\,ds}, \quad \forall t \in I, \end{aligned}$$
(59)

where \(\omega _{2}(t)=2\times 3^{q-1}(1+2^{q-1})P^{2}(t)\int _{t_{0}} ^{t}f^{q}(s)e^{-qs}\,ds\), \(p=1+\beta \), \(q=\frac{1+\beta }{\beta }\), and \(P(t)\)is defined by (11).

Proof

Using (6) and (7), from (57) we get

$$\begin{aligned} u(t) \leq & u_{0}+ \biggl[ \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}f(s)e ^{-s}u(s)\,ds \biggr]^{2} \\ & {} + \int _{t_{0}}^{t}(t-s)^{\beta -1}e^{s}f(s)e^{-s}u(s) \biggl[u(s)+2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )u(\lambda )\,d\lambda \biggr]\,ds \\ \leq & u_{0}+ \biggl[ \biggl(\frac{e^{pt}}{p^{1+p(\beta -1)}} \varGamma \bigl(1+p( \beta -1) \bigr) \biggr)^{\frac{1}{p}} \biggl( \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \biggr)^{\frac{1}{q}} \biggr]^{2} \\ & {} + \biggl(\frac{e^{pt}}{p^{1+p(\beta -1)}}\varGamma \bigl(1+p(\beta -1) \bigr) \biggr)^{\frac{1}{p}} \biggl\{ \int _{t_{0}}^{t}f^{q}(s)e^{-qs}u ^{q}(s) \biggl[u(s) \\ & {} +2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )u(\lambda )\,d \lambda \biggr]^{q}\,ds \biggr\} ^{\frac{1}{q}}, \quad \forall t\in I. \end{aligned}$$
(60)

Using (8) and from (60) we have

$$\begin{aligned} u^{q}(t) \leq &3^{q-1}u_{0}^{q}+3^{q-1}P^{2}(t) \biggl[ \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \biggr]^{2} \\ & {} +3^{q-1}P(t) \int _{t_{0}}^{t}f^{q}(s)e^{-qs}u^{q}(s) \biggl[u(s) \\ & {} +2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )u(\lambda )\,d \lambda \biggr]^{q}\,ds,\quad \forall t\in I, \end{aligned}$$
(61)

where \(P(t)\) is defined by (11).

Also using (6), (7) and (8), \(\forall t\in I\), we have

$$\begin{aligned}& \biggl[u(s)+2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )u( \lambda )\,d\lambda \biggr]^{q} \\& \quad \leq 2^{q-1}u^{q}(s)+2\times 2^{q-1}P(s) \int _{t_{0}}^{s}f^{q}(\lambda )e^{-q\lambda }u^{q}(\lambda )\,d\lambda . \end{aligned}$$
(62)

Substituting (62) into (61) we get

$$\begin{aligned} u^{q}(t) \leq &3^{q-1}u_{0}^{q}+3^{q-1}P^{2}(t) \biggl[ \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \biggr]^{2}+6^{q-1}P(t) \\ & {} \times \int _{t_{0}}^{t}f^{q}(s)e^{-qs}u^{q}(s) \biggl[u^{q}(s)+2P(s) \int _{t_{0}}^{s}f^{q}(\lambda )e^{-q\lambda }u^{q}(\lambda )\,d\lambda \biggr]\,ds \\ \leq &3^{q-1}u_{0}^{q}+3^{q-1}P^{2}(T) \biggl[ \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \biggr]^{2}+6^{q-1}P(T) \\ & {} \times \int _{t_{0}}^{t}f^{q}(s)e^{-qs}u^{q}(s) \biggl[u^{q}(s)+2P(T) \int _{t_{0}}^{s}f^{q}(\lambda )e^{-q\lambda }u^{q}(\lambda )\,d\lambda \biggr]\,ds, \end{aligned}$$
(63)

for all \(t\in [t_{0}, T]\), where \(t_{0}< T<\infty \) is chosen arbitrarily.

Let \(z_{8}(t)\) equal the right hand side in (63), we have \(z_{8}(t_{0})=3^{q-1}u_{0}^{q}\) and

$$\begin{aligned} u(t)\leq z_{8}^{\frac{1}{q}}(t). \end{aligned}$$
(64)

Differentiating \(z_{8}(t)\) with respect to t, and using (64) we obtain

$$\begin{aligned} \frac{dz_{8}(t)}{dt} =&2\times 3^{q-1}P^{2}(T)f^{q}(t)e^{-qt}u^{q}(t) \int _{t_{0}}^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \\ & {} +6^{q-1}P(T)f^{q}(t)e^{-qt}u^{q}(t) \biggl[u^{q}(t)+2P(T) \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}u^{q}(s) \,ds \biggr] \\ \leq & f^{q}(t)e^{-qt}z_{8}(t) \biggl[6^{q-1}P(T)z_{8}(t)+2\times 3^{q-1} \bigl(1+2^{q-1}\bigr)P ^{2}(T) \\ & {} \times \int _{t_{0}}^{t}f^{q}(s)e^{-qs}z_{8}(s) \,ds \biggr] \\ =&f^{q}(t)e^{-qt}z_{8}(t)Y(t), \quad \forall t\in [t_{0}, T], \end{aligned}$$
(65)

where \(Y(t)=6^{q-1}P(T)z_{8}(t)+2\times 3^{q-1}(1+2^{q-1})P^{2}(T) \int _{t_{0}}^{t}f^{q}(s)e^{-qs}z_{8}(s)\,ds\), hence \(Y(t_{0})=18^{q-1}P(T)u _{0}^{q}\), and \(z_{8}(t)\leq Y(t)\).

Differentiating \(Y(t)\) with respect to t and using (65) we obtain

$$\begin{aligned} \frac{dY(t)}{dt} \leq &6^{q-1}P(T)f^{q}(t)e^{-qt}Y^{2}(t) \\ & {} +2\times 3^{q-1}\bigl(1+2^{q-1}\bigr)P^{2}(T)f^{q}(t)e^{-qt}Y(t), \quad \forall t \in [t_{0}, T]. \end{aligned}$$
(66)

Since \(Y(t)>0\), we have

$$\begin{aligned} Y^{-2}(t)\frac{dY(t)}{dt} \leq &2\times 3^{q-1} \bigl(1+2^{q-1}\bigr)P^{2}(T)f ^{q}(t)e^{-qt}Y^{-1}(t) \\ & {} +6^{q-1}P(T)f^{q}(t)e^{-qt}, \quad \forall t \in [t_{0}, T]. \end{aligned}$$
(67)

If we let

$$\begin{aligned} S_{2}(t)=Y^{-1}(t), \quad \forall t\in [t_{0}, T], \end{aligned}$$
(68)

then we get \(S_{2}(t_{0})=Y^{-1}(t_{0})= (18^{q-1}P(T)u_{0}^{q} )^{-1}\) and \(Y^{-2}(t)\frac{dY(t)}{dt}=-\frac{dS_{2}(t)}{dt}\), thus from (67) we obtain

$$\begin{aligned} \frac{dS_{2}(t)}{dt} \geq &-2\times 3^{q-1}\bigl(1+2^{q-1} \bigr)P^{2}(T)f^{q}(t)e ^{-qt}S_{2}(t) \\ & {} -6^{q-1}P(T)f^{q}(t)e^{-qt}, \quad \forall t \in [t_{0}, T]. \end{aligned}$$
(69)

The above inequality implies the following estimation for \(S_{2}(t)\):

$$\begin{aligned} S_{2}(t)\geq \frac{1-108^{q-1}P^{2}(T)u_{0}^{q}\int _{t_{0}}^{t}f^{q}(s)e ^{-qs}\exp (c_{2}(s) )\,ds}{18^{q-1}P(T)u_{0}^{q}\exp (c_{2}(t) )}, \quad \forall t\in [t_{0}, T], \end{aligned}$$

where \(c_{2}(t)=2\times 3^{q-1}(1+2^{q-1})P^{2}(T)\int _{t_{0}}^{t}f ^{q}(s)e^{-qs}\,ds\).

Let \(t=T\), because \(t_{0}< T<\infty \) is chosen arbitrarily, then we get

$$\begin{aligned} S_{2}(t)\geq \frac{1-108^{q-1}P^{2}(t)u_{0}^{q}\int _{t_{0}}^{t}f^{q}(s)e ^{-qs}\exp (\omega _{2}(s) )\,ds}{18^{q-1}P(t)u_{0}^{q}\exp (\omega _{2}(t) )}, \quad \forall t\in I, \end{aligned}$$
(70)

then from (68) and (70) we obtain

$$\begin{aligned} Y(t)\leq \frac{18^{q-1}P(t)u_{0}^{q}\exp (\omega _{2}(t) )}{1-108^{q-1}P ^{2}(t)u_{0}^{q}\int _{t_{0}}^{t}f^{q}(s)e^{-qs} [\exp (\omega _{2}(s) ) ]\,ds}=B_{2}(t), \quad \forall t\in I, \end{aligned}$$

where \(B_{2}(t)\) is defined by (59), thus from (65) we have

$$ \frac{dz_{8}(t)}{dt}\leq f^{q}(t)e^{-qt}z_{8}(t)B_{2}(t), \quad \forall t \in I. $$

The above inequality implies the following estimation for \(z_{8}(t)\):

$$\begin{aligned} z_{8}(t)\leq 3^{q-1}u_{0}^{q}\exp \biggl( \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}B_{2}(s)\,ds \biggr). \end{aligned}$$
(71)

Then from (64) we get (58). □

Application

In this section, we present two applications of our results to the estimation of unknown functions of the integral equations.

As an application of the inequality given in Theorem 3.1, we consider the following integral equation:

$$\begin{aligned} x(t) =&x_{0}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}F_{1} \biggl(s, x(s), \int _{t_{0}}^{s}(s-\tau )^{\beta -1}F_{2} \bigl(\tau , x(\tau ) \bigr)\,d \tau \biggr)\,ds, \\ &\forall t\in I, \end{aligned}$$
(72)

where \(x_{0}\) is a positive constant. We assume that \(F_{1}\in C ([t_{0}, \infty )\times R^{2}, R )\), \(F_{2}\in C ([t_{0}, \infty )\times R, R )\) satisfy the following conditions:

$$\begin{aligned}& \bigl\vert F_{1}(t, x, y) \bigr\vert \leq f(t) \vert x \vert \bigl[ \vert x \vert + \vert y \vert \bigr]^{r}, \end{aligned}$$
(73)
$$\begin{aligned}& \bigl\vert F_{2}(t, x) \bigr\vert \leq g(t) \vert x \vert , \end{aligned}$$
(74)

where f, g are nonnegative and nondecreasing real-valued continuous functions defined on I.

Theorem 4.1

Consider the integral equation (72) and suppose that \(F_{1}\), \(F_{2}\)satisfy the conditions (73) and (74), andf, gare nonnegative and nondecreasing real-valued continuous functions defined on I. Then

$$\begin{aligned}& \begin{gathered} x(t)\leq 2^{\frac{q-1}{q}} x_{0} \biggl[\exp \biggl( \int _{t_{0}}^{t}2^{(q-1)(r+1)}P(t)f ^{q}(s)e^{-qs}B_{3}(s)\,ds \biggr) \biggr]^{\frac{1}{q}}, \quad \forall t\in I, \\ 1-2^{(q-1)(r+1)}\bigl(2^{q-1}x_{0}^{q} \bigr)^{r}rP(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs}\exp \biggl(rP(t) \int _{t_{0}}^{s}g^{q}(\tau )e^{-q\tau } \,d\tau \biggr)\,ds< 1, \end{gathered} \end{aligned}$$
(75)

where \(P(t)\)is defined by (11), \(p=1+\beta \), \(q=\frac{1+\beta }{ \beta }\)and

$$ B_{3}(t)=\frac{(2^{q-1}x_{0}^{q})^{r}\exp [rP(t)\int _{t_{0}}^{t}g^{q}(s)e ^{-qs}\,ds]}{1-2^{(q-1)(r+1)}(2^{q-1}x_{0}^{q})^{r}rP(t)\int _{t_{0}} ^{t}f^{q}(s)e^{-qs}\exp [rP(t)\int _{t_{0}}^{s}g^{q}(\tau )e^{-q\tau }\,d \tau ]\,ds}. $$

Proof

Substituting (73) and (74) into (72), we get

$$\begin{aligned} \bigl\vert x(t) \bigr\vert \leq & \vert x_{0} \vert + \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s) \bigl\vert x(s) \bigr\vert \biggl[ \bigl\vert x(s) \bigr\vert \\ & {} + \int _{t_{0}}^{s}(s-\tau )^{\beta -1}g(\tau ) \bigl\vert x(\tau ) \bigr\vert \,d \tau \biggr]^{r}\,ds, \quad \forall t \in I. \end{aligned}$$
(76)

Obviously, (76) satisfies the conditions of Theorem 3.1 and is of the form of (9). Applying Theorem 3.1 to (76), we can get the estimation (75).

As an application of the inequality given in Theorem 3.4, we consider the following integral equation:

$$\begin{aligned} x(t) = &x_{0}+ \biggl[ \int _{t_{0}}^{t}(t-s)^{\beta -1}F_{2} \bigl(s, x(s) \bigr)\,ds \biggr]^{2}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}F_{1} \biggl(s, x(s), \\ & \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}F_{2} \bigl(\lambda , x(\lambda ) \bigr)\,d\lambda \biggr)\,ds, \quad \forall t\in I, \end{aligned}$$
(77)

where \(x_{0}\) is a positive constant. We assume that \(F_{1}\in C ([t_{0}, \infty )\times R^{2}, R )\), \(F_{2}\in C ([t_{0}, \infty )\times R, R )\) satisfy the following conditions:

$$\begin{aligned}& \bigl\vert F_{1}(t, x, y) \bigr\vert \leq f(t) \vert x \vert \bigl[ \vert x \vert +2 \vert y \vert \bigr], \end{aligned}$$
(78)
$$\begin{aligned}& \bigl\vert F_{2}(t, x) \bigr\vert \leq f(t) \vert x \vert , \end{aligned}$$
(79)

where f is a nonnegative real-valued continuous functions defined on I. □

Theorem 4.2

Consider the integral equation (77) and suppose that \(F_{1}\), \(F_{2}\)satisfy the conditions (78) and (79), andfis a nonnegative and nondecreasing real-valued continuous function defined on I. Then

$$\begin{aligned}& \begin{gathered} x(t)\leq 3^{\frac{q-1}{q}}x_{0} \exp \biggl[\frac{1}{q} \int _{t_{0}} ^{t}f^{q}(s)e^{-qs}B_{4}(s) \,ds \biggr], \quad \forall t\in I, \\ 108^{q-1}P^{2}(t)x_{0}^{q} \int _{t_{0}}^{t}f^{q}(s)e^{-qs} \bigl[\exp \bigl(\omega _{2}(s) \bigr) \bigr]\,ds< 1, \end{gathered} \end{aligned}$$
(80)

where \(P(t)\)is defined by (11), \(p=1+\beta \), \(q=\frac{1+\beta }{ \beta }\)and

$$\begin{aligned}& B_{4}(t)=\frac{18^{q-1}P(t)x_{0}^{q}\exp (\omega _{2}(t) )}{1-108^{q-1}P ^{2}(t)x_{0}^{q}\int _{t_{0}}^{t}f^{q}(s)e^{-qs} [\exp (\omega _{2}(s) ) ]\,ds}, \\& \omega _{2}(t)=2\times 3^{q-1}\bigl(1+2^{q-1} \bigr)P^{2}(t) \int _{t_{0}}^{t}f^{q}(s)e ^{-qs} \,ds. \end{aligned}$$

Proof

Substituting (78) and (79) into (77), we get

$$\begin{aligned} \bigl\vert x(t) \bigr\vert \leq & \vert x_{0} \vert + \biggl[ \int _{t_{0}}^{t}(t-s)^{ \beta -1}f(s)x(s)\,ds \biggr]^{2}+ \int _{t_{0}}^{t}(t-s)^{\beta -1}f(s)x(s) \biggl[x(s) \\ & {} +2 \int _{t_{0}}^{s}(s-\lambda )^{\beta -1}f(\lambda )x(\lambda )\,d \lambda \biggr]\,ds, \quad \forall t\in I. \end{aligned}$$
(81)

Obviously, (81) satisfies the conditions of Theorem 3.4 and is of the form of (57). Applying Theorem 3.4 to (81), we can get the estimation (80). □

References

  1. 1.

    Gronwall, H.T.: Note on the derivatives with respect to a parameter of the solutions of a system of differential equations. Ann. Math. 20, 292–296 (1919)

  2. 2.

    Bellman, R.: The stability of solutions of linear differential equations. Duke Math. J. 10, 643–647 (1943)

  3. 3.

    Jiang, F.C., Meng, F.W.: Explicit bounds on some new nonlinear integral inequalities with delay. J. Comput. Appl. Math. 205, 479–486 (2007)

  4. 4.

    Feng, Q.H., Meng, F.W., Zheng, B.: Gronwall–Bellman type nonlinear delay integral inequalities on time scales. J. Math. Anal. Appl. 382, 772–784 (2011)

  5. 5.

    Meng, F.W., Li, W.N.: On some new integral inequalities and their applications. J. Comput. Appl. Math. 148, 381–392 (2004)

  6. 6.

    Pachpatte, B.G.: Inequalities for Differential and Integral Equations. Academic Press, New York (1998)

  7. 7.

    Xu, R., Meng, F.W.: Some new weakly singular integral inequalities and their applications to fractional differential equations. J. Inequal. Appl. 2016, 78 (2016)

  8. 8.

    Ma, Q.H., Pecaire, J.: Some new explicit bounds for weakly singular integral inequalities with applications to fractional differential and integral equations. J. Math. Anal. Appl. 341, 894–905 (2008)

  9. 9.

    Abdeldaim, A., Yakout, M.: On some new integral inequalities of Gronwall–Belman–Pachpatte type. Appl. Math. Comput. 217, 7887–7899 (2001)

  10. 10.

    Wang, W.S.: Some new generalized retarded nonlinear integral inequalities with iterated integrals and their applications. J. Inequal. Appl. 2012, 236 (2012)

  11. 11.

    Wang, W.S., Huang, C.M.: Estimation of unknown function of weakly singular iterated integral inequality. J. South China Normal Univ. Natur. Sci. Ed. 49, 111–114 (2017)

  12. 12.

    Medved, M.: A new approach to an analysis of Henry type integral inequalities and their Bihari type versions. J. Math. Anal. Appl. 214, 349–366 (1997)

  13. 13.

    Kuczma, M.: An Introduction to the Theory of Functional Equations and Inequalities: Cauchy’s Equation and Jensen’s Inequality. University of Katowice (1985)

Download references

Acknowledgements

The authors are very grateful to the anonymous referees for their valuables suggestions and comments, which helped to improve the quality for the paper.

Funding

This research is supported by National Science Foundation of China (No. 11671227) and the Natural Science Foundation of Shandong Province (No. ZR2019MA034).

Author information

The main idea of this paper was proposed by YR and MFW. YR prepared the manuscript initially and performed all the steps of the proofs in this research. All authors read and approved the final manuscript.

Correspondence to Fanwei Meng.

Ethics declarations

Competing interests

The authors declare that there is no conflict of interest regarding the publication of this paper.

Additional information

Publisher’s Note

Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.

Rights and permissions

Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article’s Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article’s Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creativecommons.org/licenses/by/4.0/.

Reprints and Permissions

About this article

Verify currency and authenticity via CrossMark

Cite this article

Yan, R., Meng, F. Estimation of unknown function of nonlinear weakly singular integral inequalities of Gronwall–Bellman–Pachpatte type. J Inequal Appl 2019, 319 (2019). https://doi.org/10.1186/s13660-019-2269-2

Download citation

Keywords

  • Weakly singular integral inequalities
  • Discrete Jensen inequality
  • Hölder integral inequality
  • Ordinary differential equation