Page 125 - IJOCTA-15-1
P. 125

Exponential stability for higher-order impulsive fractional neutral stochastic integro-delay . . .

            refers the complete orthogonal basis in H and
                     ∞ √                                          n
                     P                                         p−1  p−1 ¯ p     p   p   p−1  p  p  p      p  p
            w(ς) =       λ n w n (ς)Z n , ς ≥ 0, λ n > 0, where  6  2  M (1 + M ) l + 2    M ς L (1 + M ) l
                                                                                             ς
                                                                         g
                                                                                                  f
                                                                                σ 1
                                                                                                         σ 2
                    n=1
            {w n (ς)} n≥0 are one-dimensional standard Bm.       p−1   p ˜   p  p  p  p−1  p p  ∞   1   p q
                                                                                                X
                                                              + 2    M C p ς 2 L l + 2   M l       L k
                                                                       ς
            Define the operator A : D(A) ⊂ H → H by A =                         G          ϑ
             ∂  2                       2          1                                            k=1
             ∂ς 2 , with domain D(A) = H ([0, π]) ∩ H ([0, π]),  p−1  p p  ∞  2   p o
                                                                                  q
                                                                           X
            where                                             + 2   M l       L k    < 1.
                                                                      ς
                                                                           k=1
                                                                    n                                p
                                                                               p
                                                                                                p ˜
                                                                                        p
                                                                                      p
                                                                      ¯ p
                                                              12 p−1 p  M (1 + M ) + M ς ˆa 3 + M C p ς 2 ˆa 5
                                                                   l
                                                                                               ς
                                                                        g
                                                                                      ς
                                         ∂C                                    σ 1
                                                2
                1
                                 2
              H ([0, π]) = {C ∈ L ([0, π]),  ∈ L ([0, π]),       ˆ a 7 π 2  ˆ a 8 π  4 o
                                         ∂Z                   +      +       < 1.
                         C(0) = C(π) = 0},                        6     90
                                         ∂C                   Then, we can easily obtain the result that the sys-
                                 2
                2
                                                2
              H ([0, π]) = {C ∈ L ([0, π]),  ∈ L ([0, π])}.   tem (12) has a mild solution. It is evident that
                                         ∂Z
                                                              all the assumptions are satisfied with a 1 = a 2 =
                                                                               1
                                                                                         2
            Then                                              1, β 1 = β 2 = π, L =  ˆ a 7 2 , L =  ˆ a 8 4 ,
                                                                                         k
                                                                               k
                                                                                             k
                                                                                    k
                                                                   n                    p  1−p    p
                                                                                p
                                                                     ¯ p
                                                                                                           p
                          ∞                                   8 p−1  M (1 + M ) + [a β         [L (1 + M )]
                                                                       g
                         X      2                                               σ 1     2  2      f       σ 2
                   Ax =      −n (x, Z n ),  x ∈ D(A),                    1−  p
                                                                  p
                                                                     ˜
                                                                                                      L
                                                                             L G ] + a β
                         n=1                                  + a C p β  2  2 ¯      p   2 −p CHς pH−1 ˆ
                                                                                     2
                                                                  2
                                                                    ∞       p      ∞
                            q                                        X        q       X       p o
                                                                                               q
            where Z n (ξ) =    2  sin(nξ) is the orthonormal  + a p      L 1   + a p      L 2     < 1.
                              π                                   1        k       2        k
            set of eigenfunctions of A. A is the infinitesimal       k=1              k=1
            generator of a strongly continuous cosine family   p−1 n       p 1−p        p 1−  p
                                                              8     ˆ a 1 + a β  ˆ a 3 + a β  2 ˜
                                                                                              C p ˆa 5 + ˆa 6
            {C(ς), ς ∈ R}, define as                                       2 2          2 2
                                                                  p  ˆ a 7 π 2  p  ˆ a 8 π 4 o
                                                              + a         + a          < 1.
                            ∞                                     1   6       2  90
                           X
                  C(ξ)Z =     cosnξ < (x, Z n ), Z ∈ H,
                                                              Thus, all the assumptions (H 1 )−(H 7 ) and (H 9 ) of
                           n=1
                                                              Theorem 2 are satisfied. So we conclude that the
                                                              mild solution of the system (12) is exponentially
            and the associated sine family is given by
                                                              stable.
                         ∞                                    6. Conclusion
                        X    1
               S(ξ)Z =     − sinnξ < (x, Z n ) >, Z ∈ H.
                             n                                The existence and uniqueness results have been
                        n=1
                                                              investigated for a class of higher-order impul-
                                                              sive FNSIDDEs with nonlocal conditions driven
            Next, we give a special B-space.    Let l(ϑ) =
                                  0                           by mixed fBm with infinite delays in a Hilbert
                                  R
             2ϑ
            e , ϑ < 0 then, l 0 =   l(ϑ)dϑ =  1  and define   space. The well-posedness of mild solution for
                                              2
                                 −∞                           (1) is achieved by using the BFPT along with
                                                              theories of the semigroup of linear operators and
                           0
                          Z                                   stochastic processes. Some novel sufficient con-
                                                 1
                                               p
                  ∥φ∥ B =    l(ϑ) sup (E∥φ(θ)∥ ) dϑ.          ditions have been derived by using BFPT. Next,
                                                 p
                                 ϑ≤θ≤0
                         −∞                                   the exponential stability of mild solution higher-
                                                              order impulsive FNSIDDEs with nonlocal con-
            It follows from Hino, 36  that (B b , ∥·∥ B b ) is a Banach  ditions driven by mixed fBm in stochastic set-
            space. Thus, for (ς, φ) ∈ J × B b , where φ(θ)(x) =  tings has been obtained by the integral inequality
            φ(θ, x)(θ, x) ∈ (−∞, 0]×[0, π], Z(ς)(x) = Z(ς, x).  technique. Furthermore, it was shown that sys-
                                                              tem is well-posed even if the system (1) is inte-
                                                              grated along with the nonlocal conditions. Mov-
            Accordingly, all the assumptions of Theorem 1 are  ing ahead in this direction, exponential stability
            satisfied and in particular, ς =  1 , M ϑ = M ς =  of non-instantaneous impulsive neutral stochastic
                                            2
                                             1
                                                ˜
                    = 0.01, M g = 0.02, α = , C p = 1, l =    integro-differential equations of Sobolev type in-
                                             4
            1, M σ 1
            0.5, ˆa 1 = 0.075, ˆa 2 = 0.003, ˆa 3 = 0.001, ˆa 5 =  volving Poisson jumps can be considered in the
            0.025, ˆa 7 = 0.002, ˆa 8 = 0.03, p = 2;          future. The numerical example presented can be
                                                           119
   120   121   122   123   124   125   126   127   128   129   130