
X.-M. Zhang, Q.-L. Han, X. Ge et al. Automatica 147 (2023) 110592
Table 1
The admissible upper bound ℓ2for different values of ℓ1.
ℓ15 7 11 13 NoDVs
Nam et al. (2015) 20 22 26 27 29.5n2+12.5n
Chen et al. (2016) 21 22 26 27 78.5n2+12.5n
Zhang et al. (2016) 21 22 26 27 10.5n2+3.5n
Chen et al. (2020) 21 23 27 29 39.5n2+6.5n
Proposition 3 (d=5) 27 29 33 34 341.5n2+8.5n
Proposition 3 (d=2) 28 30 33 35 341.5n2+8.5n
Ωi(ℓ)=ℓ¯
Φ1i+¯
Φ0i1
2ℓ¯
Φ2i
1
2ℓ¯
Φ2iℓ¯
Φ3i,i=1,2,
Ω0(ℓ)=ℓ¯
Φ10 +¯
Φ00 1
2ℓ¯
Φ20
1
2ℓ¯
Φ20 ℓ¯
Φ30 +He{M1[ℓI−I]}.
3.3. Comparison on numerical complexities
From the analysis above, two stability criteria, Propositions 3
and 5, are obtained using a convex method and a polynomial
inequality method, respectively. In the following, we make a
comparison between their numerical complexities.
By Han (2009), the numerical complexity of Proposition 3 or
5is proportional to LM3, where Lis the total row size of the
related matrix inequalities, and Mmeans the total number of
scalar decision variables. We now calculate the pairs (L1,M1)
and (L2,M2) of the total row size Ljand the total number
Mj(j=1,2) of scalar decision variables for Propositions 3 and
5, respectively.
•For Proposition 3:L1=100n,M1=341.5n2+8.5n;
•For Proposition 5:L2=192n,M2=1372.5n2+8.5n.
Thus, the numerical complexity of Proposition 5 is propor-
tional to L2M3
2=192n×(1372.5n2+8.5n)3, which is much
bigger than L1M3
1=100n×(341.5n2+8.5n)3.
For a small number of n=2, the numerical complexity of
Proposition 5 is proportional to L2M3
2=384 ×55073>1.5×
212 ×1010, which is extremely high. Therefore, Proposition 5 is
unworkable to check if the system (1) is asymptotically stable
while Proposition 3 is applicable. In Section 4, through simulation,
one can see that Proposition 3 can deliver larger delay upper
bounds than some existing ones. However, how to reduce the
numerical complexity of Proposition 3 is still challenging.
Remark 6. In this paper, we develop a convex method to study
the stability of delayed discrete-time systems by introducing a
novel Lyapunov functional (3) with a matrix P(k)=ℓ2(k)P2+
ℓ(k)P1+P0. The matrix P(k) can be extended to a general one
like P(k)=s
i=0ℓi(k)Pi, where sis a positive integer. Similar to
the proof of Proposition 3, one can get a stability criterion based
on the Lyapunov functional with P(k). Nevertheless, the resultant
stability criterion should be of higher numerical complexity for a
larger s.
4. Numerical examples
Consider the system (1), where
A=0.648 0.04
0.12 0.654,B=−0.1512 −0.0518
0.0259 −0.1091(31)
This example is widely studied in the literature. For different
values of ℓ1, the admissible upper bounds of ℓ2are listed in
Table 1 by applying some existing stability criteria, e.g. Chen et al.
(2016, Theorem 1), Chen et al. (2020, Theorem 1), Nam et al.
(2015, Theorem1) and Zhang et al. (2016, Theorem1). It should
be pointed out that, those stability criteria are independent of
the delay-variation rate. Thus, they may be conservative when
information on the delay-variation rate is available, that is, when
the bounds d1and d2are known. In fact, if d=d2= −d1∈
{2,5}for this example, Proposition 3 can deliver larger upper
bound ℓ2for the same values of ℓ1, which can be seen in Table 1.
However, the required number of decision variables (NoDVs) in
Proposition 3 is much bigger than those in Chen et al. (2016,
2020), Nam et al. (2015) and Zhang et al. (2016).
On the other hand, the matrix Bin this example is nonsin-
gular. If Bis singular, Proposition 3 is still available to check
the stability of the system. In fact, if we change the matrix B
in (31) to [−0.1512 −0.0518
0 0 ], then for d=2, the admissible upper
bounds of ℓ2are computed as 27,29,32,34 for ℓ1=5,7,11,13,
respectively.
5. Conclusion
Stability of discrete-time systems with time-varying delay has
been investigated in this paper by introducing a novel Lyapunov
functional V(k) with a quadratic polynomial matrix. An effective
method to deal with the forward difference of V(k) has been
presented, leading to a less conservative stability criterion, which
has been illustrated through a numerical example.
References
Chen, J., Lu, J., & Xu, S. (2016). Summation inequality and its application to
stability analysis for time-delay systems. IET Control Theory & Applications,
10(4), 391–395.
Chen, J., Park, J., Xu, S., & Zhang, X.-M. (2020). Stability of discrete-time
systems with time-varying delay via a novel Lyapunov-Krasovskii functional.
International Journal of Robust and Nonlinear Control,30, 4779–4788.
Gu, K., Kharitonov, V. L., & Chen, J. (2003). Stability of time-delay systems.
Birkhäuser.
Han, Q.-L. (2009). A discrete delay decomposition approach to stability of linear
retarded and neutral systems. Automatica,45, 1948–1952.
Kao, C. (2012). On stability of discrete-time LTI systems with varying time delays.
IEEE Transactions on Automatic Control,57(5), 1243–1248.
Kwon, O., Park, M., Park, J., Lee, S., & Cha, E. (2012). Improved robust stability
criteria for discrete-time systems with interval time-varying delays via new
zero equalities. IET Control Theory & Applications,6(16), 2567–2575.
Nam, P. T., Trinh, H., & Pathirana, P. N. (2015). Discrete inequalities based on
multiple auxiliary functions and their applications to stability analysis of
time-delay systems. Journal of the Franklin Institute,352, 5810–5831.
Seuret, A., & Gouaisbaut, F. (2013). Wirtinger-based integral inequality:
Application to time-delay systems. Automatica,49, 2860–2866.
Seuret, A., Gouaisbaut, F., & Fridman, E. (2015). Stability of discrete-time systems
with time-varying delays via a novel summation inequality. IEEE Transactions
on Automatic Control,60(10), 2740–2745.
Xu, S., Lam, J., & Zhou, Y. (2005). Improved conditions for delay-dependent
robust stability and stabilization of uncertain discrete time-delay systems.
Asian Journal of Control,7(3), 344–348.
Zhang, X.-M., & Han, Q.-L. (2015). Abel lemma-based finite-sum inequality and
its application to stability analysis for linear discrete time-delay systems.
Automatica,57, 199–202.
Zhang, X.-M., Han, Q.-L., & Ge, X. (2021). Sufficient conditions for a class of
matrix-valued polynomial inequalities on closed intervals and application to
H∞filtering for linear systems with time-varying delays. Automatica,125,
Article 109390.
Zhang, X.-M., Han, Q.-L., & Ge, X. (2022). A novel approach to H∞performance
analysis of discrete-time networked systems subject to network-induced
delays and malicious packet dropouts. Automatica,136, Article 110010.
Zhang, C.-K., He, Y., Jiang, L., & Wu, M. (2016). An improved summation
inequality to discrete-time systems with time-varying delay. Automatica,74,
10–15.
5