Abstract
This paper is concerned with an impulsive non-autonomous high-order Hopfield neural network with mixed delays. Under proper conditions, we studied the existence, the uniqueness and the global exponential stability of asymptotic almost automorphic solutions for the suggested system. Our method was mainly based on the Banach’s fixed-point theorem and the generalized Gronwall–Bellman inequality. Moreover, four examples are presented to demonstrate the effectiveness of the proposed findings.
Similar content being viewed by others
References
Abbas S, Mahto L, Hafayed M, Alimi AM (2014) Asymptotic almost automorphic solutions of impulsive neural network with almost automorphic coefficients. Neurocomputing 142:326–33
Abbas S, Kavitha V, Murugesu R (2015) Stepanov-like weighted pseudo almost automorphic solutions to fractional order abstract integro-differential equations. Proc Math Sci 125(3):323–351
Abbas S, Chang YK, Hafayed M (2014) Stepanov type weighted pseudo almost automorphic sequences and their applications to difference equations. Nonlinear Stud 21(1):99–111
Abbas S, Yonghui XIA (2013) Existence and attractivity of k-almost automorphic sequence solution of a model of cellular neural networks with delay. Acta Math Sci 33(1):290–302
Abbas S, Xia Y (2015) Almost automorphic solutions of impulsive cellular neural networks with piecewise constant argument. Neural Process Lett 42(3):691–702
Ammar B, Chérif F, Alimi AM (2012) Existence and uniqueness of pseudo almost-periodic solutions of recurrent neural networks with time-varying coefficients and mixed delays. IEEE Trans Neural Netw Learn Syst 23(1):109–118
Aouiti C, M’hamdi MS, Touati A (2017) Pseudo almost automorphic solutions of recurrent neural networks with time-varying coefficients and mixed delays. Neural Process Lett 45(1):121–140
Aouiti C, M’hamdi MS, Cao J, Alsaedi A (2017) Piecewise pseudo almost periodic solution for impulsive generalised high-order Hopfield neural networks with leakage delays. Neural Process Lett 45(2):615–648
Aouiti C (2016) Oscillation of impulsive neutral delay generalized high-order Hopfield neural networks. Neural Comput Appl. https://doi.org/10.1007/s00521-016-2558-3
Aouiti C (2016) Neutral impulsive shunting inhibitory cellular neural networks with time-varying coefficients and leakage delays. Cogn Neurodyn 10(6):573–591
Aouiti C, M’hamdi MS, Chérif F (2017) New results for impulsive recurrent neural networks with time-varying coefficients and mixed delays. Neural Process Lett 46(2):487–506
Aouiti C, Coirault P, Miaadi F, Moulay E (2017) Finite time boundedness of neutral high-order Hopfield neural networks with time delay in the leakage term and mixed time delays. Neurocomputing 260:378–392
Aouiti C, M’hamdi MS, Chérif F, Alimi AM (2017) Impulsive generalised high-order recurrent neural networks with mixed delays: stability and periodicity. Neurocomputing. https://doi.org/10.1016/j.neucom.2017.11.037
Brahmi H, Ammar B, Chérif F, Alimi AM (2014) On the dynamics of the high-order type of neural networks with time varying coefficients and mixed delay. In: 2014 international joint conference on neural networks (IJCNN), pp 2063–2070
Brahmi H, Ammar B, Chérif F, Alimi AM, Abraham A (2016) Asymptotically almost automorphic solution of high order recurrent neural networks with mixed delays. Int J Comput Sci Inf Secur 14(7):284
Brahmi H, Ammar B, Alimi AM, Chérif F (2016) Pseudo almost periodic solutions of impulsive recurrent neural networks with mixed delays. In: 2016 international joint conference on neural networks (IJCNN), pp 464–470. IEEE
Cai SM, Xu FD, Zheng WX, Liu ZR (2009) Exponential stability analysis for impulsive neural networks with time-varying delays. In: Optimization and systems biology: the third international symposium, OSB’09, Zhangjiajie, China, 20–22 Sept 2009. Proceedings, pp 81–88
Cao J, Wang L (2002) Exponential stability and periodic oscillatory solution in BAM networks with delays. IEEE Trans Neural Netw 13(2):457–463
Cao J (2003) New results concerning exponential stability and periodic solutions of delayed cellular neural networks. Phys Lett A 307(2):136–147
Cao J, Wang J (2005) Global asymptotic and robust stability of recurrent neural networks with time delays. IEEE Trans Circuits Syst I Regul Pap 52(2):417–426
Cao J, Liang J, Lam J (2004) Exponential stability of high-order bidirectional associative memory neural networks with time delays. Phys D Nonlinear Phenom 199(3):425–436
Cao J, Chen A, Huang X (2005) Almost periodic attractor of delayed neural networks with variable coefficients. Phys Lett A 340(1):104–120
Cao J, Song Q (2006) Stability in Cohen–Grossberg-type bidirectional associative memory neural networks with time-varying delays. Nonlinearity 19(7):1601–1617
Chang YK, Cheng ZX, N’Guérékata GM (2016) Stepanov-like pseudo almost automorphic solutions to some stochastic differential equations. Bull Malays Math Sci Soc 39(1):181–197
Chang YK, Luo XX (2015) Pseudo almost automorphic behavior of solutions to a semi-linear fractional differential equation. Math Commun 20(1):53–68
Chang YK, Bian YT (2015) Weighted asymptotic behavior of solutions to a Sobolev-type differential equation with Stepanov coefficients in Banach spaces. Filomat 29(6):1315–1328
Chang YK, Zhang R, N’Guérékata GM (2014) Weighted pseudo almost automorphic solutions to nonautonomous semilinear evolution equations with delay and ${S}^{p} $-weighted pseudo almost automorphic coefficients. Topol Methods Nonlinear Anal 43(1):69–88
Chang YK, Luo XX (2014) Existence of $\mu $-pseudo almost automorphic solutions to a neutral differential equation by interpolation theory. Filomat 28(3):603–614
Chavez A, Castillo S, Pinto M (2013) Discontinuous almost automorphic functions and almost automorphic solutions of differential equations with piecewise constant argument. Electron J Differ Equ 56:113
Chen A, Cao J (2003) Existence and attractivity of almost periodic solutions for cellular neural networks with distributed delays and variable coefficients. Appl Math Comput 134(1):125–140
Chérif F (2014) Sufficient conditions for global stability and existence of almost automorphic solution of a class of RNNs. Differ Equ Dyn Syst 22(2):191–207
Diagana T (2013) Almost automorphic type and almost periodic type functions in abstract spaces. Springer, New York
Gerlee P, Anderson AR (2009) Modelling evolutionary cell behaviour using neural networks: application to tumour growth. Biosystems 95(2):166–174
Huang X, Cao J, Ho DW (2006) Existence and attractivity of almost periodic solution for recurrent neural networks with unbounded delays and variable coefficients. Nonlinear Dyn 45(3):337–351
Kavitha V, Abbas S, Murugesu R (2015) Asymptotically almost automorphic solutions of fractional order neutral integro-differential equations. Bull Malays Math Sci Soc 39(3):1075–1088
Kavitha V, Wang PZ, Murugesu R (2013) Existence of weighted pseudo almost automorphic mild solutions to fractional integro-differential equations. J Fract Calc Appl 4(1):37–55
Li Y (2013) Periodic solutions of non-autonomous cellular neural networks with impulses and delays on time scales. IMA J Math Control Inf 31(2):273–293
Liang J, Zhang J, Xiao TJ (2008) Composition of pseudo almost automorphic and asymptotically almost automorphic functions. J Math Anal Appl 340(2):1493–1499
M’hamdi MS, Aouiti C, Touati A, Alimi AM, Snasel V (2016) Weighted pseudo almost-periodic solutions of shunting inhibitory cellular neural networks with mixed delays. Acta Math Sci 36(6):1662–1682
Mahto L, Abbas S (2015) PC-almost automorphic solution of impulsive fractional differential equations. Mediterr J Math 12(3):771–790
Marcus CM, Westervelt RM (1988) Dynamics of analog neural networks with time delay. In: NIPS, pp 568–576
Marcus CM, Westervelt RM (1989) Stability of analog neural networks with delay. Phys Rev A 39(1):347
Moghtadaei M, Golpayegani MRH, Malekzadeh R (2013) A variable structure fuzzy neural network model of squamous dysplasia and esophageal squamous cell carcinoma based on a global chaotic optimization algorithm. J Theor Biol 318:164–172
Mohamad S (2007) Exponential stability in Hopfield-type neural networks with impulses. Chaos Solitons Fractals 32(2):456–467
Mohamad S, Gopalsamy K, Akca H (2008) Exponential stability of artificial neural networks with distributed delays and large impulses. Nonlinear Anal Real World Appl 9(3):872–888
N’Guérékata GM (1981) Sur les solutions presque automorphes d’équations différentielles abstraites. Ann Sci Math Quebec 5(1):69–79
N’Guérékata GM (1987) Some remarks on asymptotically almost automorphic functions. Riv Math Universita di Parma 13(4):301–303
Rakkiyappan R, Pradeep C, Vinodkumar A, Rihan FA (2013) Dynamic analysis for high-order Hopfield neural networks with leakage delay and impulsive effects. Neural Comput Appl 22(1):55–73
Ren F, Cao J (2006) LMI-based criteria for stability of high-order neural networks with time-varying delay. Nonlinear Anal Real World Appl 7(5):967–979
Ren F, Cao J (2007) Periodic oscillation of higher-order bidirectional associative memory neural networks with periodic coefficients and delays. Nonlinearity 20(3):605–629
Ren F, Cao J (2007) Periodic solutions for a class of higher-order Cohen–Grossberg type neural networks with delays. Comput Math Appl 54(6):826–839
Stamov GT (2012) Almost periodic solutions of impulsive differential equations. Springer, Berlin
Stamov GT (2004) Impulsive cellular neural networks and almost periodicity. Proc Jpn Acad Ser A Math Sci 80(10):198–203
Shi P, Dong L (2010) Existence and exponential stability of anti-periodic solutions of Hopfield neural networks with impulses. Appl Math Comput 216(2):623–630
Tonnesen J (2013) Optogenetic cell control in experimental models of neurological disorders. Behav Brain Res 255:35–43
Tyagi S, Abbas S, Hafayed M (2016) Global Mittag–Leffler stability of complex valued fractional-order neural network with discrete and distributed delays. Rendiconti del Circolo Matematico di Palermo Series 2 65(3):485–505
Wang C (2016) Piecewise pseudo almost periodic solution for impulsive non-autonomous high-order Hopfield neural networks with variable delays. Neurocomputing 171:1291–1301
Wang C, Agarwal RP (2015) Weighted piecewise pseudo almost automorphic functions with applications to abstract impulsive $\nabla $-dynamic equations on time scales. Adv Differ Equ 2014(1):153
Wang J, Jiang H, Hu C (2014) Existence and stability of periodic solutions of discrete-time Cohen–Grossberg neural networks with delays and impulses. Neurocomputing 142:542–550
Wang Y, Xiong W, Zhou Q, Xiao B, Yu Y (2006) Global exponential stability of cellular neural networks with continuously distributed delays and impulses. Phys Lett A 350(1):89–95
Weng YF, Ju L, Wang J (2007) Cellular neural networks and biological visual information processing model. J Beijing Technol Bus Univ (Nat Sci Ed) 25(1):42–58
Xiong W (2015) New result on convergence for HCNNs with time-varying leakage delays. Neural Comput Appl 26(2):485–491
Xu C, Li P (2016) Pseudo almost periodic solutions for high-order Hopfield neural networks with time-varying leakage delays. Neural Process Lett 46(1):41–58
Xu B, Liu X, Teo KL (2009) Asymptotic stability of impulsive high-order Hopfield type neural networks. Comput Math Appl 57(11):1968–1977
Xia Z (2016) Pseudo almost periodic mild solution of nonautonomous impulsive integro-differential equations. Mediterr J Math 13(3):1065–1086
Yang Y, Cao J (2007) Stability and periodicity in delayed cellular neural networks with impulsive effects. Nonlinear Anal Real World Appl 8(1):362–374
Yang X, Cao J, Huang C, Long Y (2010) Existence and global exponential stability of almost periodic solutions for SICNNs with nonlinear behaved functions and mixed delays. In: Abstract and applied analysis. Hindawi Publishing Corporation
Zhang Q, Wei X, Xu J (2003) Global exponential stability of Hopfield neural networks with continuously distributed delays. Phys Lett A 315(6):431–436
Zhu Q, Liang F, Zhang Q (2009) Global exponential stability of Cohen-Grossberg neural networks with time-varying delays and impulses. J Shanghai Univ (Engl Ed) 13(3):255–259
Author information
Authors and Affiliations
Corresponding author
Ethics declarations
Conflict of interest
There is no conflict of interest.
Appendices
Appendix 1: Proof of the Lemma 1
Proof
Let \(\varphi (.)\in AAA(\mathbb {R},\mathbb {R})\), it can be written as \(\varphi (.)=\varphi _1 (.)+\varphi _2 (.)\) where \(\varphi _1 (.) \in AA(\mathbb {R},\mathbb {R})\) and \(\varphi _2 (.)\in PC_0(\mathbb {R},\mathbb {R}).\)
First, we know that the space \(AA(\mathbb {R},\mathbb {R})\) is translation invariant, then for \(h\in \mathbb {R},\) we have \(\varphi _1(.-h) \in AA(\mathbb {R},\mathbb {R}).\)
Second, we prove that \(\varphi _2(.-h) \in PC_0(\mathbb {R},\mathbb {R}).\)
For \(\varphi _2 (.)\in PC_0(\mathbb {R},\mathbb {R}),\) we have: \(\varphi _2 (.) \in PC(\mathbb {R},\mathbb {R}),\) such that \(\varphi _2(t)\) is continuous at t for any \(t \notin \{ t_i , i \in \mathbb {Z}\},\)\(\varphi _2(t_i^+),\varphi _2(t_i^-)\) exists and \(\varphi _2(t_i^-)=\varphi _2(t_i).\)
Therefore, for \(h\in \mathbb {R},\)\(\varphi _2(t-h)\) is continuous at \((t-h)\) for any \((t-h)\notin \{ t_i , i \in \mathbb {Z}\},\)\(\varphi _2((t_i-h)^+),\varphi _2((t_i-h)^-)\) exist and \(\varphi _2((t_i-h)^-)=\varphi _2(t_i-h).\) Then, \(\varphi _2(t-h) \in PC(\mathbb {R},\mathbb {R}).\)
On the other hand, we have \(\lim \nolimits _{t\longrightarrow \infty } \Vert \varphi _2(t)\Vert =0, \;\) then for h in \(\mathbb {R},\lim \nolimits _{t\longrightarrow \infty } \Vert \varphi _2(t-h)\Vert =0.\) This completes the proof. \(\square\)
Appendix 2: Proof of Lemma 2
Proof
By definition, we can write \(\varphi =\varphi _1+\varphi _2,\)\(\psi =\psi _1+\psi _2\) where \(\varphi _1,\psi _1 \in AA(\mathbb {R},\mathbb {R}),\)\(\varphi _2,\psi _2 \in PC_0(\mathbb {R},\mathbb {R}).\)
Obviously, \(\varphi \times \psi =\varphi _1 \times \psi _1 +\varphi _1 \times \psi _2 +\varphi _2 \times \psi _1 +\varphi _2 \times \psi _2,\) we have \(\varphi _1 \times \psi _1 \in AA(\mathbb {R},\mathbb {R}).\)
On the other hand, \(\varphi _1 \times \psi _2 +\varphi _2 \times \psi _1 +\varphi _2 \times \psi _2 \in PC(\mathbb {R},\mathbb {R}),\) and
which implies that \(\varphi _1 \times \psi _2 +\varphi _2 \times \psi _1 +\varphi _2 \times \psi _2 \in PC_0(\mathbb {R},\mathbb {R}).\)
Then, \(\varphi \times \psi \in AAA(\mathbb {R},\mathbb {R}).\) This completes the proof. \(\square\)
Appendix 3: Proof of Lemma 3
Proof
By definition, we have \(\phi (.)=\phi _1 (.)+\phi _2 (.)\) where \(\phi _1 (.)\in AA(\mathbb {R},\mathbb {R}^n), \phi _2 (.)\in PC_0(\mathbb {R},\mathbb {R}^n).\) Let
First, let \(\left( s_{n}^{\prime }\right) _{n \in \mathbb {N}}\) be a sequence of real numbers. By hypothesis we can extract a subsequence \((s_{n})_{n \in \mathbb {N}}\) of \(\left( s_{n}^{\prime }\right) _{n \in \mathbb {N}}\) such that \(\lim \nolimits _{n\rightarrow +\infty } \phi _1\left( t-\varsigma +s_{n}\right) =\phi _1^1(t-\varsigma ),\; \forall t \in \mathbb {R}\) and \(\lim \nolimits _{n\rightarrow +\infty } \phi _1^1\left( t-\varsigma -s_{n}\right) =\phi _1(t-\varsigma ),\; \forall t \in \mathbb {R}.\) Obviously,
Therefore, \(\lim \nolimits _{t\longrightarrow \infty }f(\phi _1(t-\varsigma +s_{n}))=f(\phi _1^1(t-\varsigma )).\)
By the same way, we have: \(\lim \nolimits _{t\longrightarrow \infty }f(\phi _1^1(t-\varsigma -s_{n}))=f(\phi _1(t-\varsigma )).\)
Then \(G_1(.) \in AA(\mathbb {R},\mathbb {R}^n).\)
Second, we prove that \(G_2(.)\in PC_0(\mathbb {R},\mathbb {R}^n)\).
It is clear that \(G_2(.)\in PC(\mathbb {R},\mathbb {R}^n)\), also we have:
since \(\phi _2(.) \in PC_0(\mathbb {R},\mathbb {R}^n),\) we have \(\lim \nolimits _{t\longrightarrow \infty } |\phi _2(t-\varsigma )|=0,\) then \(G_2(.)\in PC_0(\mathbb {R},\mathbb {R}^n).\) The proof is completed. \(\square\)
Appendix 4: Proof of Lemma 4
Proof
Let \(\phi _{j}(.) \in AAA(\mathbb {R},\mathbb {R}^n),\) from Lemma 3 we obtain \(h_{j}(\phi _{j}(.) )\in AAA(\mathbb {R},\mathbb {R}^n).\)
Let \(h_{j}(\phi _{j}(.))=u_j(.)+v_j(.),\) where \(u_j (.)\in AA(\mathbb {R},\mathbb {R}^n)\) and \(v_j (.)\in PC_0(\mathbb {R},\mathbb {R}^n),\) then
First, let us show that \(\varPhi _{ij}^1(t) \in AA(\mathbb {R},\mathbb {R}^n).\)
For each sequence \((s'_n)\) there exists a subsequence \((s_n)\) such that \(\theta (t)= \lim \nolimits _{n\longrightarrow \infty } u_j(t+s_n)\) is well defined for every \(t \in \mathbb {R}\) and \(\theta (t-s_n)= \lim \nolimits _{n\longrightarrow \infty } u_j(t)\) is well defined for every \(t \in \mathbb {R}.\)
In addition, we have
One has \(\Vert K_{ij}(t-s) u_{j}(s+s_n)\Vert \le K^+ e^{-\nu ^K(t-s)}\Vert u_{j}(t)\Vert\) it follows that \(\int \nolimits _{-\infty }^{t} K_{ij}(t-s) u_{j}(s+s_n) \,{\text{d}}s\le \frac{K^+}{\nu ^K}\Vert u_{j}(t)\Vert .\)
Then using the Lebesgue-dominated convergence theorem, we obtain \(\lim \nolimits _{n\longrightarrow \infty } \varPhi _{ij}^1(t+s_n)=\int \nolimits _{-\infty }^{t} K_{ij}(t-s) \theta _j(s)\,{\text{d}}s.\)
Analogously, we get \(\varPhi _{ij}^1(t)=\lim \nolimits _{n\longrightarrow \infty } \int \nolimits _{-\infty }^{t-s_n} K_{ij}(t-s_n-s) \theta _j(s)\,{\text{d}}s\).
Second, let us show that \(\varPhi _{ij}^2(t) \in PC_0(\mathbb {R},\mathbb {R}^n).\)
It is not difficult to see that \(\varPhi _{ij}^2(t) \in PC(\mathbb {R},\mathbb {R}^n).\) We have
since \(v_j \in PC_0(\mathbb {R},\mathbb {R}^n),\) for every \(\varepsilon >0\) there exist a constant \(N>0\) such that \(\Vert v_j (s)\Vert \le \varepsilon\) for all \(s\ge N\) and for all \(t\ge 2N\), we obtain
where \(\Vert v_j\Vert _\infty = \sup \nolimits _{s \in \mathbb {R}}\Vert v_j(s)\Vert .\)
Consequently \(\varPhi _{ij}^2(.) \in PC_0(\mathbb {R},\mathbb {R}^n).\) The proof is completed. \(\square\)
Appendix 5: Proof of Lemma 6
Proof
Step 1 Noting \((\varPsi _{U_{\phi }})_{i}(s):=\int \nolimits _{-\infty }^{t} W(t,s)(U_\phi )_i(s)\,{\text{d}}s.\)
First, by Lemmas 1–4, the function \((U_{\phi })_{i}\) belongs to \(AAA(\mathbb {R},\mathbb {R}).\) This ensures the existence of two functions \(\varLambda _{i}\) in \(AA(\mathbb {R},\mathbb {R})\) and \(\varOmega _{i}\) in \(PC_0(\mathbb {R},\mathbb {R})\) such that for all \(1\le i,j\le n,\) it can be expressed as \((U_{\phi })_{i}(.)=\varLambda _{i}(.)+\varOmega _{i}(.).\)
One can write \(\varPsi\) as follows:
\((\varPsi _{U_{\phi }})_i(t):=\int \limits _{-\infty }^{t} W(t,s)\varLambda _{i}(s)\,{\text{d}}s+\int \limits _{-\infty }^{t} W(t,s) \varOmega _{i}(s)\,{\text{d}}s.\)
Let us study the almost automorphicity of
\((\varPsi \varLambda _{i}): t\mapsto \int \limits _{-\infty }^{t} W(t,s)\varLambda _{i}(s) \,{\text{d}}s.\)
Let \(\left( s_{n}^{\prime }\right) _{n \in \mathbb {N}}\) be a sequence of real numbers. By hypothesis we can extract a subsequence \((s_{n})_{n \in \mathbb {N}}\) of \(\left( s_{n}^{\prime }\right) _{n \in \mathbb {N}}\) such that: \(\lim \nolimits _{n\rightarrow +\infty }\varLambda _{i}\left( t+s_{n}\right) =\varLambda _{i}^{1}\left( t\right) ,\)\(\forall \; t\in \mathbb {R},\)
and \(\lim \nolimits _{n\rightarrow +\infty }\varLambda _{i}^{1}\left( t-s_{n}\right) =\varLambda _{i}\left( t\right) ,\)\(\forall \; t\in \mathbb {R}.\)
Let \((\varPsi ^1\varLambda _{i})(t)= \int \nolimits _{-\infty }^{t} W(t,s)\varLambda _{i}^1(s) \,{\text{d}}s,\) it follows that
Based on the Lebesgue-dominated convergence theorem, we have for all \(t \in \mathbb {R}\)
By a similar way, we prove that
which implies that \((\varPsi \varLambda _{i}) \in AA(\mathbb {R},\mathbb {R}^{n}).\)
Second, we turn our attention to \((\varPsi \varOmega _{i}): t\mapsto \int \limits _{-\infty }^{t} W(t,s)\varOmega _{i}(s) \,{\text{d}}s.\) It is easy to prove that \((\varPsi \varOmega _{i})\in PC(\mathbb {R},\mathbb {R}).\)
We have \(\lim \nolimits _{t\rightarrow +\infty } \int \nolimits _{-\infty }^{t} W(t,s) \varOmega _{i}(s)\,{\text{d}}s=0.\) Since \(\varOmega _{i} \in PC_0(\mathbb {R},\mathbb {R}),\) then \(\lim \nolimits _{t\rightarrow +\infty } |\int \nolimits _{-\infty }^{t} W(t,s) \varOmega _{i}(s)|\,{\text{d}}s=0.\)
By the Lebesgue-dominated convergence theorem, we have
\(\lim \limits _{t\rightarrow +\infty } \int \limits _{-\infty }^{t} W(t,s) \varOmega _{i}(s)\,{\text{d}}s=0.\)
Hence, the function \(\varPsi \varOmega _{i}\) belongs to \(PC_0(\mathbb {R},\mathbb {R}).\)
Step 2 Proving that \(\sum \limits _{t_k<t} W(t,t_k)(I_k(\phi _i(t_k))+\omega _k)\) belongs to \(AAA(\mathbb {R},\mathbb {R}).\)
From the assumption (H7), \(I_k(\phi _i(t_k))\in AAA(\mathbb {R},\mathbb {R}).\) By definition, it can be expressed as
such that \(I_{k1}(\phi _i(t_k)) \in AA(\mathbb {R},\mathbb {R}),\)\(I_{k2}(\phi _i(t_k))=0.\) Then:
For every real sequence \((t_{n})_{n\in \mathbb {N}}\), there exists a subsequence \((t_{n_{k}})_{n_{k} \in \mathbb {N}}\) such that \(\lim \limits _{n_k\rightarrow +\infty }I_{k1}(\phi _i(t_k+t_{n_{k}}))=I_{k1}^1(\phi _i(t_k))\) and
\(\lim \limits _{n_k\rightarrow +\infty }I_{k1}^1(\phi _i(t_k-t_{n_{k}}))=I_{k1}(\phi _i(t_k)).\)
Now, we have
then
Similarly
then
Then, \(\sum \limits _{t_k<t} W(t,t_k)(I_{k1}(\phi _i(t_k))+\omega _k) \in AA(\mathbb {R},\mathbb {R}).\)
On the other hand,
as \(\sum \nolimits _{t_k< t} |W(t,t_k)|<\infty .\)
By Steps 1 and 2 we have:
maps \(AAA(\mathbb {R},\mathbb {R})\) into itself. \(\square\)
Appendix 6: Proof of the Theorem 1
Proof
Let us calculate the norm of \(\phi _0\). One has
such that \(\bar{\gamma }\ge \max \bigg \{\max \limits _{1 \le i \le n}|\gamma _i(t)|, \max \limits _{1 \le k \le n} |\omega _k|\bigg \}.\)
After, \(\Vert \phi \Vert _{\infty }\le \Vert \phi -\phi _{0}\Vert +\Vert \phi _{0}\Vert \le \frac{r}{1-r}\bar{R}+\bar{R }=\frac{\bar{R}}{1-r}.\)
Set \(S^{*}=\bigg \{ \phi \in AAA(\mathbb {R},\mathbb {R}^{n}) ; \Vert \phi -\phi _{0}\Vert \le \frac{r}{1-r}\bar{R} \bigg \}.\)
Clearly, \(S^{*}\) is a closed convex subset of \(AAA(\mathbb {R},\mathbb {R}^{n}).\) Therefore, for any \(\phi \in S^{*}\) by using the estimate just obtained, we see that
then, \(\varTheta _{\phi }\in S^{*}.\)
Now our aim is to prove that \(\varTheta\) is a contraction. For any \(\phi _{1},\phi _{2} \in S^{*},\) we have
which prove that \(\varTheta\) is a contraction mapping.
By virtue of the Banach’s fixed-point theorem, \(\varTheta\) has a unique fixed point which corresponds to the solution of (3) in \(S^{*}.\)\(\square\)
Appendix 7: Proof of Theorem 2
Proof
First, using Lemma 6, \(\varTheta\) has a fixed point \(\phi .\) Let \(I^*_k(\phi (t_k))=I_k(\phi (t_k))+\omega _k.\)
Hence, for all \(t\in \mathbb {R},\) the fixed point \(\phi\) satisfies the following integral system:
\(\phi (t):= \int \limits _{-\infty }^{t} W(t,s)U_\phi (s)\,{\text{d}}s+ \sum \limits _{t_k<t} W(t,t_k)(I^*_k(\phi (t_k))).\)
Fixed \(t_0,\)\(t_0\ne t_i,\)\(i \in \mathbb {Z},\) we have
Therefore
Second, by Theorem 1, we know that system (3) has an asymptotically almost automorphic solution u(t), by using integral form of system (3), if \(t>\sigma ,\;\; \sigma \ne t_k,\; \; k \in \mathbb {Z}\)
Let \(u(t)=u(t,\sigma ,\phi _1)\) and \(v(t)=v(t,\sigma ,\phi _2)\) be two solutions of (3), then
Therefore,
Then
Let \(y(t)=e^{\delta t}\Vert u(t)-v(t)\Vert ,\) Eq. (28) can be rewritten in the following form:
By the generalized Gronwall–Bellman inequality, we have
Since \(\vartheta =\inf \nolimits _{k\in \mathbb {Z}}(t_{k+1}-t_k)>0\), we have
That is \(\Vert u(t)-v(t)\Vert \le K\Vert \phi _1-\phi _2\Vert e^{(\zeta -\delta )(t-\sigma )}.\)
Since \((\zeta -\delta )<0,\) then system (3) has an exponential stable asymptotically almost automorphic solution. This completes the proof. \(\square\)
Rights and permissions
About this article
Cite this article
Aouiti, C., Dridi, F. Piecewise asymptotically almost automorphic solutions for impulsive non-autonomous high-order Hopfield neural networks with mixed delays. Neural Comput & Applic 31, 5527–5545 (2019). https://doi.org/10.1007/s00521-018-3378-4
Received:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s00521-018-3378-4