VMSTA Modern Stochastics: Theory and Applications 2351-60542351-60462351-6046 VTeXMokslininkų g. 2A, 08412 Vilnius, Lithuania VMSTA274 10.15559/25-VMSTA274 Research Article About stability of equilibria of one system of stochastic delay differential equations with exponential nonlinearity https://orcid.org/0000-0001-7354-1383 ShaikhetLeonidleonid.shaikhet@usa.net Department of Mathematics, Ariel University, Ariel 40700, Israel 2025 132202512331332412102024122025122025 © 2025 The Author(s). Published by VTeX2025 Open access article under the CC BY license.

A system of two nonlinear delay differential equations under stochastic perturbations is considered. Nonlinearity of the exponential type in each equation of the system under consideration depends on the both variables of the system. The stability in probability of the zero and nonzero equilibria of the system is studied via the general method of Lyapunov functionals construction and the method of linear matrix inequalities (LMIs). The obtained results are illustrated via examples and figures with numerical simulations of solutions of a considered system of stochastic differential equations. The proposed way of investigation can be applied to nonlinear systems of higher dimension and with other types of nonlinearity, both for delay differential equations and for difference equations.

 Nonlinear differential equation stochastic perturbations asymptotic mean square stability stability in probability linear matrix inequality (LMI) exponential nonlinearity 34G20 34K20 34K50 Introduction

Systems of differential and difference equations with various types of exponential nonlinearities are very popular both in theoretical research and in different applications (see, for instance, [2, 4, 5, 7–9, 15, 18–22, 24, 26–29, 33–41] and references therein).

Here, similarly to [38], the stability of the positive equilibrium of a system with exponential nonlinearity is investigated under stochastic perturbations via the general method of Lyapunov functionals construction [16, 17, 31, 32] and the method of linear matrix inequalities (LMIs) [3, 6, 10–12, 14, 23, 25, 30]. However, unlike, for instance, [28, 37–39], where the exponential nonlinearity in each equation depends on only one variable, here each equation exponentially depends on all variables of the system under consideration. The obtained results are illustrated via examples and figures with numerical simulations of solutions of the considered system of stochastic differential equations. Numerical analysis of the considered LMI is carried out using MATLAB.

Consider the system of two nonlinear delay differential equations x ˙ 1 ( t ) + a 1 x 1 ( t ) + b 1 x 1 ( t − h ) = c 1 x 1 ( t ) e − p 1 x 1 ( t ) − q 1 x 2 ( t ) , x ˙ 2 ( t ) + a 2 x 2 ( t ) + b 2 x 2 ( t − h ) = c 2 x 2 ( t ) e − p 2 x 1 ( t ) − q 2 x 2 ( t ) , x i ( s ) = ϕ i ( s ) , s ∈ [ − h , 0 ] , i = 1 , 2 , t ≥ 0 , with positive parameters and positive initial conditions.

It is clear that the system (1.1) has the zero equilibrium E 0 = ( 0 , 0 ), and the nonzero equilibrium E ∗ = ( x 1 ∗ , x 2 ∗ ) of the system (1.1) is defined by the system of two algebraic equations a 1 + b 1 = c 1 e − p 1 x 1 − q 1 x 2 , a 2 + b 2 = c 2 e − p 2 x 1 − q 2 x 2 , that can be presented in the form of the system of two linear equations p 1 x 1 + q 1 x 2 = r 1 , p 2 x 1 + q 2 x 2 = r 2 , r i = ln ( c i / ( a i + b i ) ) , i = 1 , 2 , with the solution x i ∗ = Δ i Δ , i = 1 , 2 , Δ = p 1 q 1 p 2 q 2 , Δ 1 = r 1 q 1 r 2 q 2 , Δ 2 = p 1 r 1 p 2 r 2 .

Note that via (1.3) for an equilibrium E ∗ = ( x 1 ∗ , x 2 ∗ ) with both positive coordinates the condition r i > 0, i.e. c i > a i + b i , i = 1 , 2, must be satisfied. It is clear that for c i < a i + b i we get a negative equilibrium, while for c i = a i + b i there is the zero equilibrium. However, the inequality c i > a i + b i is only a necessary condition for the solution of the system (1.3) to be positive, but not a sufficient one. Indeed, for example, the system 2 x 1 + x 2 = 4, x 1 + x 2 = 1 with r 1 = 4, r 2 = 1 has the solution x 1 = 3, x 2 = − 2, the system 2 x 1 + x 2 = 4, x 1 + x 2 = 2 with r 1 = 4, r 2 = 2 has the solution x 1 = 2, x 2 = 0.

Below, the system (1.1) is investigated under stochastic perturbations, the conditions of stability for both the zero and one positive equilibria are studied, the obtained results are illustrated by numerical simulations of solutions of the system under consideration.

 Stochastic perturbations, centralization and linearization

Let ( x 1 ∗ , x 2 ∗ ) be one of the equilibria of the system (1.1). Let us suppose that the system (1.1) is exposed to stochastic perturbations of the white noise type, that are directly proportional to the deviation of the system (1.1) state ( x 1 ( t ) , x 2 ( t ) ) from the equilibrium ( x 1 ∗ , x 2 ∗ ). As a result, we obtain the system of Ito’s stochastic delay differential equations [13] d x 1 ( t ) = ( − a 1 x 1 ( t ) − b 1 x 1 ( t − h ) + c 1 x 1 ( t ) e − p 1 x 1 ( t ) − q 1 x 2 ( t ) ) d t + σ 1 ( x 1 ( t ) − x 1 ∗ ) d w 1 ( t ) , d x 2 ( t ) = ( − a 2 x 2 ( t ) − b 2 x 2 ( t − h ) + c 2 x 2 ( t ) e − p 2 x 1 ( t ) − q 2 x 2 ( t ) ) d t + σ 2 ( x 2 ( t ) − x 2 ∗ ) d w 2 ( t ) .

Note that stochastic perturbations of the type of (2.1) were first used in [1] and later in many other works (see, for instance, [31, 32] and references therein). In this case, the equilibrium ( x 1 ∗ , x 2 ∗ ) of the deterministic system (1.1) is also the solution of the stochastic system (2.1).

Nonzero equilibrium

Let ( x 1 ∗ , x 2 ∗ ) be the nonzero equilibrium of the system (1.1). Using the new variables y i ( t ) = x i ( t ) − x i ∗ , i = 1 , 2, and (1.2), transform the first equation of the system (2.1) by the following way: d y 1 ( t ) = [ − a 1 ( y 1 ( t ) + x 1 ∗ ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) + c 1 ( y 1 ( t ) + x 1 ∗ ) e − p 1 y 1 ( t ) − q 1 y 2 ( t ) e − p 1 x 1 ∗ − q 1 x 2 ∗ ] d t + σ 1 y 1 ( t ) d w 1 ( t ) = [ − a 1 ( y 1 ( t ) + x 1 ∗ ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) + ( a 1 + b 1 ) ( y 1 ( t ) + x 1 ∗ ) e − p 1 y 1 ( t ) − q 1 y 2 ( t ) ] d t + σ 1 y 1 ( t ) d w 1 ( t ) = [ ( y 1 ( t ) + x 1 ∗ ) ( ( a 1 + b 1 ) e − p 1 y 1 ( t ) − q 1 y 2 ( t ) − a 1 ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) ] d t + σ 1 y 1 ( t ) d w 1 ( t ) . Similarly, for the second equation of the system (2.1) we have d y 2 ( t ) = [ − a 2 ( y 2 ( t ) + x 2 ∗ ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) + c 2 ( y 2 ( t ) + x 2 ∗ ) e − p 2 y 1 ( t ) − q 2 y 2 ( t ) e − p 2 x 1 ∗ − q 2 x 2 ∗ ] d t + σ 2 y 2 ( t ) d w 2 ( t ) = [ − a 2 ( y 2 ( t ) + x 2 ∗ ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) + ( a 2 + b 2 ) ( y 2 ( t ) + x 2 ∗ ) e − p 2 y 1 ( t ) − q 2 y 2 ( t ) ] d t + σ 2 y 2 ( t ) d w 2 ( t ) = [ ( y 2 ( t ) + x 2 ∗ ) ( ( a 2 + b 2 ) e − p 2 y 1 ( t ) − q 2 y 2 ( t ) − a 2 ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) ] d t + σ 2 y 2 ( t ) d w 2 ( t ) . As a result, we obtain the system of two nonlinear Ito’s stochastic delay differential equations [13] with the zero solution: d y 1 ( t ) = [ ( y 1 ( t ) + x 1 ∗ ) ( ( a 1 + b 1 ) e − p 1 y 1 ( t ) − q 1 y 2 ( t ) − a 1 ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) ] d t + σ 1 y 1 ( t ) d w 1 ( t ) , d y 2 ( t ) = [ ( y 2 ( t ) + x 2 ∗ ) ( ( a 2 + b 2 ) e − p 2 y 1 ( t ) − q 2 y 2 ( t ) − a 2 ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) ] d t + σ 2 y 2 ( t ) d w 2 ( t ) .

Using the representation e − x = 1 − x + o ( x ), where lim x → 0 o ( x ) x = 0, we obtain the linear part of the system (2.2). Really, neglecting the nonlinear terms in the square brackets of the first equation (2.2), we have ( y 1 ( t ) + x 1 ∗ ) ( ( a 1 + b 1 ) e − p 1 y 1 ( t ) − q 1 y 2 ( t ) − a 1 ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) = ( y 1 ( t ) + x 1 ∗ ) ( ( a 1 + b 1 ) ( 1 − p 1 y 1 ( t ) − q 1 y 2 ( t ) ) − a 1 ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) = ( y 1 ( t ) + x 1 ∗ ) ( b 1 − ( a 1 + b 1 ) ( p 1 y 1 ( t ) + q 1 y 2 ( t ) ) ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) = b 1 ( y 1 ( t ) + x 1 ∗ ) − ( a 1 + b 1 ) x 1 ∗ ( p 1 y 1 ( t ) + q 1 y 2 ( t ) ) − b 1 ( y 1 ( t − h ) + x 1 ∗ ) = ( b 1 − ( a 1 + b 1 ) x 1 ∗ p 1 ) y 1 ( t ) − ( a 1 + b 1 ) x 1 ∗ q 1 y 2 ( t ) − b 1 y 1 ( t − h ) . Similarly, for the second equation (2.2) we obtain ( y 2 ( t ) + x 2 ∗ ) ( ( a 2 + b 2 ) e − p 2 y 1 ( t ) − q 2 y 2 ( t ) − a 2 ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) = ( y 2 ( t ) + x 2 ∗ ) ( ( a 2 + b 2 ) ( 1 − p 2 y 1 ( t ) − q 2 y 2 ( t ) ) − a 2 ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) = ( y 2 ( t ) + x 2 ∗ ) ( b 2 − ( a 2 + b 2 ) ( p 2 y 1 ( t ) + q 2 y 2 ( t ) ) ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) = b 2 ( y 2 ( t ) + x 2 ∗ ) − ( a 2 + b 2 ) x 2 ∗ ( p 2 y 1 ( t ) + q 2 y 2 ( t ) ) − b 2 ( y 2 ( t − h ) + x 2 ∗ ) = − ( a 2 + b 2 ) x 2 ∗ p 2 y 1 ( t ) + ( b 2 − ( a 2 + b 2 ) x 2 ∗ q 2 ) y 2 ( t ) − b 2 y 2 ( t − h ) . As a result, we obtain the linear part of the system (2.2) d z 1 ( t ) = [ ( b 1 − ( a 1 + b 1 ) x 1 ∗ p 1 ) z 1 ( t ) − ( a 1 + b 1 ) x 1 ∗ q 1 z 2 ( t ) − b 1 z 1 ( t − h ) ] d t + σ 1 z 1 ( t ) d w 1 ( t ) , d z 2 ( t ) = [ − ( a 2 + b 2 ) x 2 ∗ p 2 z 1 ( t ) + ( b 2 − ( a 2 + b 2 ) x 2 ∗ q 2 ) z 2 ( t ) − b 2 z 2 ( t − h ) ] d t + σ 2 z 2 ( t ) d w 2 ( t ) , or in the matrix form d z ( t ) = ( A z ( t ) − B z ( t − h ) ) d t + ∑ i = 1 2 C i z ( t ) d w i ( t ) , where z ( t ) = z 1 ( t ) z 2 ( t ) , A = b 1 − ( a 1 + b 1 ) x 1 ∗ p 1 − ( a 1 + b 1 ) x 1 ∗ q 1 − ( a 2 + b 2 ) x 2 ∗ p 2 b 2 − ( a 2 + b 2 ) x 2 ∗ q 2 , B = b 1 0 0 b 2 , C 1 = σ 1 0 0 0 , C 2 = 0 0 0 σ 2 .

Zero equilibrium

Let us linearize the system (2.1) with the zero equilibrium. Neglecting the nonlinear terms, transform the expression in brackets of the first equation (2.1) by the following way: − a 1 x 1 ( t ) − b 1 x 1 ( t − h ) + c 1 x 1 ( t ) e − p 1 x 1 ( t ) − q 1 x 2 ( t ) = − a 1 x 1 ( t ) − b 1 x 1 ( t − h ) + c 1 x 1 ( t ) ( 1 − p 1 x 1 ( t ) − q 1 x 2 ( t ) ) = ( c 1 − a 1 ) x 1 ( t ) − b 1 x 1 ( t − h ) . Similarly, for the second equation (2.1) we have − a 2 x 2 ( t ) − b 2 x 2 ( t − h ) + c 2 x 2 ( t ) e − p 2 x 1 ( t ) − q 2 x 2 ( t ) = − a 2 x 2 ( t ) − b 2 x 2 ( t − h ) + c 2 x 2 ( t ) ( 1 − p 2 x 1 ( t ) − q 2 x 2 ( t ) ) = ( c 2 − a 2 ) x 2 ( t ) − b 2 x 2 ( t − h ) .

As a result, we obtain that the linearized system decomposes into two unrelated equations of the same type d z 1 ( t ) = [ ( c 1 − a 1 ) z 1 ( t ) − b 1 z 1 ( t − h ) ] d t + σ 1 z 1 ( t ) d w 1 ( t ) , d z 2 ( t ) = [ ( c 2 − a 2 ) z 2 ( t ) − b 2 z 2 ( t − h ) ] d t + σ 2 z 2 ( t ) d w 2 ( t ) .

Stability Some necessary definitions and statements

Let { Ω , F , P } be a complete probability space, { F t } t ≥ 0 be a nondecreasing family of sub-σ-algebras of F, i.e. F s ⊂ F t for s < t, P { · } be the probability of an event enclosed in the braces, E be the mathematical expectation, H 2 be the space of F 0 -adapted stochastic processes φ ( s ) = ( φ 1 ( s ) , φ 2 ( s ) ), s ∈ [ − h , 0 ], ‖ φ ‖ 0 = sup s ∈ [ − h , 0 ] | φ ( s ) |, ‖ φ ‖ 1 2 = sup s ∈ [ − h , 0 ] E | φ ( s ) | 2 , | φ ( s ) | = φ 1 2 ( s ) + φ 2 2 ( s ) , | φ ( s ) | 2 = φ 1 2 ( s ) + φ 2 2 ( s ).

The zero solution of the system (2.2) with the initial condition y ( s ) = ( y 1 ( s ) , y 2 ( s ) ) = ϕ ( s ), s ∈ [ − h , 0 ], is called stable in probability if for any ε 1 > 0 and ε 2 > 0 there exists δ > 0 such that the solution y ( t , ϕ ) of the system (2.2) satisfies the condition P { sup t ≥ 0 | y ( t , ϕ ) | > ε 1 } < ε 2 for any initial function ϕ such that P { ‖ ϕ ‖ 0 < δ } = 1.

The zero solution of the equation (2.3) with the initial condition z ( s ) = ϕ ( s ), s ∈ [ − h , 0 ], is called:

mean square stable if for each ε > 0 there exists a δ > 0 such that E | z ( t , ϕ ) | 2 < ε, t ≥ 0, provided that ‖ ϕ ‖ 1 2 < δ;

asymptotically mean square stable if it is mean square stable and lim t → ∞ E | z ( t , ϕ ) | 2 = 0 for each initial function ϕ.

Note that the level of nonlinearity of the system (2.2) is more than one. It is known [32] that in this case sufficient conditions for asymptotic mean square stability of the zero solution of the linear part of this system, i.e. of the equation (2.3), are also sufficient conditions for stability in probability of the zero solution of the nonlinear system (2.2) and therefore are sufficient conditions for stability in probability of the equilibrium ( x 1 ∗ , x 2 ∗ ) of the system (2.1).

Let z ( t ) be a value of the solution of the equation (2.3) in the time moment t and z t = z ( t + s ), s ∈ [ − h , 0 ], be a trajectory of the solution of the equation (2.3) until the time moment t.

Consider a functional V ( t , φ ) : [ 0 , ∞ ) × H 2 → R + that can be represented in the form V ( t , φ ) = V ( t , φ ( 0 ) , φ ( s ) ), s ∈ [ − h , 0 ), and for φ = z t put V φ ( t , z ) = V ( t , φ ) = V ( t , z t ) = V ( t , z , z ( t + s ) ) , z = φ ( 0 ) = z ( t ) , s ∈ [ − h , 0 ) .

Let D be the set of the functionals, for which the function V φ ( t , z ) defined by (3.1) has a continuous derivative with respect to t and two continuous derivatives with respect to z. The generator L of the equation (2.3) is defined on the functionals from D and has the form [13, 32] (here and everywhere below ′ is the sign of transposition) L V ( t , z t ) = ∂ ∂ t V φ ( t , z ( t ) ) + ∇ V φ ′ ( t , z ( t ) ) ( A z ( t ) − B x ( t − h ) ) + 1 2 ∑ i = 1 2 z ′ ( t ) C i ′ ∇ 2 V φ ( t , z ( t ) ) C i z ( t ) .

([<xref ref-type="bibr" rid="j_vmsta274_ref_032">32</xref>]).

Let there exist a functional V ( t , φ ) ∈ D, positive constants c 1 , c 2 , c 3 , such that the following conditions hold: E V ( t , x t ) ≥ c 1 E | x ( t ) | 2 , E V ( 0 , ϕ ) ≤ c 2 ‖ ϕ ‖ 2 , E L V ( t , x t ) ≤ − c 3 E | x ( t ) | 2 . Then the zero solution of the equation (2.3) is asymptotically mean square stable.

 Nonzero equilibrium

Let there exist positive definite matrices P and R, such that the linear matrix inequality (LMI) Φ = P A + A ′ P + S 0 + R − P B − B ′ P − R < 0 holds, where via (2.4) S 0 = ∑ i = 1 2 C i ′ P C i = σ 1 2 p 11 0 0 σ 2 2 p 22 and p 11 , p 22 are diagonal elements of the matrix P. Then the equilibrium ( x 1 ∗ , x 2 ∗ ) of the system (2.1) is stable in probability.

Following Remark 3.1, it is enough to prove the asymptotic mean square stability of the zero solution of the linear equation (2.3). Using the general method of Lyapunov functionals construction [16, 17, 31, 32], consider the functional V = V 1 + V 2 , where V 1 ( z ( t ) ) = z ′ ( t ) P z ( t ), P > 0, the additional functional V 2 will be chosen below. Via (2.3), (3.2) and ∇ V 1 ( z ( t ) = 2 P z ( t ), ∇ 2 V 1 ( z ( t ) ) = 2 P, we have L V 1 ( z ( t ) ) = 2 z ′ ( t ) P ( A z ( t ) − B x ( t − h ) ) + z ′ ( t ) S 0 z ( t ) = z ′ ( t ) ( P A + A ′ P + S 0 ) z ( t ) − 2 z ′ ( t ) P B z ( t − h ) . Putting V 2 ( z t ) = ∫ t − h t z ′ ( s ) R z ( s ) d s, R > 0, with L V 2 ( z t ) = z ′ ( t ) R z ( t ) − z ′ ( t − h ) R z ( t − h ), as a result for the functional V = V 1 + V 2 , we obtain L V ( z ( t ) ) = z ′ ( t ) ( P A + A ′ P + S 0 + R ) z ( t ) − 2 z ′ ( t ) P B z ( t − h ) − z ′ ( t − h ) R z ( t − h ) = η ′ ( t ) Φ η ( t ) , where η ( t ) = ( z ( t ) , z ( t − h ) ) ′ and the matrix Φ is defined in (3.3). From here and the LMI (3.3) it follows that L V ( z ( t ) ) ≤ − c | η ( t ) | 2 ≤ − c | z ( t ) | 2 for some c > 0. Via Theorem 3.1 it means that the zero solution of the linear equation (2.3) is asymptotically mean square stable.

Via Remark 3.1 the sufficient conditions for asymptotic mean square stability of the zero solution of the linear equation (2.3) are also sufficient conditions for stability in probability of the zero solution of the nonlinear system (2.2) and therefore are sufficient conditions for stability in probability of the equilibrium ( x 1 ∗ , x 2 ∗ ) of the system (2.1). The proof is completed.  □

Consider the system (2.1) with the following values of the parameters: a 1 = 1 , b 1 = 0.02 , c 1 = 4 , p 1 = 2 , q 1 = 0.2 , σ 1 = 1 , h = 1 , a 2 = 2 , b 2 = 0.03 , c 2 = 5 , p 2 = 0.2 , q 2 = 2 , σ 2 = 1 . By that from (1.4) and (2.4) we obtain x 1 ∗ = 0.6446 , x 2 ∗ = 0.3862 , A = − 1.2950 − 0.1315 − 0.1568 − 1.5381 , B = 0.02 0 0 0.03 , C 1 = 1 0 0 0 , C 2 = 0 0 0 1 . Via MATLAB it was shown that for the parameters (3.4), (3.5) there exist the positive definite matrices P = 301.6380 − 23.3803 − 23.3803 252.7107 , R = 268.6902 4.3515 4.3515 281.0097 , for which the LMI (3.3) holds. It means that the equilibrium ( x 1 ∗ , x 2 ∗ ) = ( 0.6446 , 0.3862 ) of the system (2.1) is stable in probability. In Fig. 1 100 trajectories of the solution of the system (2.1) are shown, with the initial conditions x 1 ( s ) = 0.77, x 2 ( s ) = 0.27, s ∈ [ − h , 0 ]. One can see that all trajectories converge to the equilibrium.

100 trajectories of the solution of the system (2.1): x 1 ( t ) (blue) and x 2 ( t ) (red)

Graph showing blue decreasing and red increasing curves over time, converging towards 0.6 and 0.4 respectively from 0 to 6 on x-axis. Zero equilibrium

Note that both equations of the system (2.5) have the form x ˙ ( t ) = a x ( t ) + b x ( t − h ) + σ x ( t ) w ˙ ( t ) , where a = c i − a i , b = − b i , σ, h ≥ 0 are known constants.

([<xref ref-type="bibr" rid="j_vmsta274_ref_032">32</xref>]).

The necessary and sufficient condition for asymptotic mean square stability of the zero solution of the equation (3.6) is a + b < 0 , G − 1 > 1 2 σ 2 , where G = b q − 1 sin ( q h ) − 1 a + b cos ( q h ) , b + | a | < 0 , q = b 2 − a 2 , 1 + | a | h 2 | a | , b = a < 0 , b q − 1 sinh ( q h ) − 1 a + b cosh ( q h ) , a + | b | < 0 , q = a 2 − b 2 . If, in particular, b = 0, then this necessary and sufficient stability condition takes the form 2 a + σ 2 < 0.

Via Remark 3.1 and Lemma 3.1, we obtain

Let the conditions c i < a i + b i , G i − 1 > 1 2 σ i 2 , i = 1 , 2 , hold, where G i = 1 + b i q i − 1 sin ( q i h ) a i − c i + b i cos ( q i h ) , | a i − c i | < b i , q i = b i 2 − ( a i − c i ) 2 , 1 + ( a i − c i ) h 2 ( a i − c i ) , b i = a i − c i > 0 , 1 + b i q i − 1 sinh ( q i h ) a i − c i + b i cosh ( q i h ) , b i < a i − c i , q i = ( a i − c i ) 2 − b i 2 . Then the zero equilibrium ( x 1 ∗ , x 2 ∗ ) = ( 0 , 0 ) of the system (2.1) is stable in probability.

Via the equation (3.6), for the proof it is enough to use Lemma 3.1, putting into (3.7) and (3.8) a = c i − a i , b = − b i , G = G i , σ = σ i . By that, (3.7) and (3.8) are transformed into (3.9) and (3.10). The proof is completed.  □

100 trajectories of the solution of the system (2.1): x 1 ( t ) (blue) and x 2 ( t ) (red)

Graph showing multiple overlapping decay curves in blue and red, with x-axis from 0 to 3 and y-axis from 0 to 1, peaking near origin.

Consider the system (2.1) with the following values of the parameters: a 1 = 2 , b 1 = 1 , c 1 = 1 , p 1 = 2 , q 1 = 0.2 , σ 1 = 0.9 , h = 1 , a 2 = 3 , b 2 = 1 , c 2 = 2 , p 2 = 0.2 , q 2 = 2 , σ 2 = 0.9. It is easy to see, that for the values of the parameters (3.11) the conditions (3.9) hold. Therefore, the zero equilibrium of the system (2.1) is stable in probability. In Fig. 2 100 trajectories of the solution of the system (2.1) are shown, with the initial conditions x 1 ( s ) = 0.55, x 2 ( s ) = 0.35, s ∈ [ − h , 0 ]. One can see that all trajectories converge to the zero.

 Conclusions

A system of two differential equations with exponential nonlinearity is considered. The nonlinearity in each equation depends on both variables of the system. It is supposed that this system is exposed to stochastic perturbations of the white noise type, which are directly proportional to the deviation of the system state from one of the system’s equilibria (the zero or nonzero). Some sufficient conditions for stability in probability of the both system equilibria are obtained by virtue of the general method of Lyapunov functionals construction and the method of linear matrix inequalities (LMIs). The obtained results are illustrated by numerical simulations of solutions of the considered system. The method of stability investigation described here can be applied to systems of higher dimension and with other types of nonlinearity.

 References Beretta, E., Kolmanovskii, V., Shaikhet, L.: Stability of epidemic model with time delays influenced by stochastic perturbations. Math. Comput. Simul. 45(3-4), 269–277 (1998). Special Issue “Delay Systems”. MR1622405. https://doi.org/10.1016/S0378-4754(97)00106-7 Berezansky, L., Braverman, E., Idels, L.: Nicholson’s blowflies differential equations revisited: Main results and open problems. Appl. Math. Model. 34(6), 1405–1417 (2010). MR2592579. https://doi.org/10.1016/j.apm.2009.08.027 Boyd, S., El-Ghaoui, L., Feron, E., Balakrishnan, V.: Linear Matrix Inequalities in System and Control Theory. SIAM, Philadelphia, PA, USA (1994). MR1284712. https://doi.org/10.1137/1.9781611970777 Bradul, N., Shaikhet, L.: Stability of the positive point of equilibrium of Nicholson’s blowflies equation with stochastic perturbations: numerical analysis. Discrete Dyn. Nat. Soc. 25, 92959 (2007). MR2346516. https://doi.org/10.1155/2007/92959 Chen, W., Wang, W.: Global exponential stability for a delay differential neoclassical growth model. Adv. Differ. Equ. 2014(9), 325 (2014). MR3360574. https://doi.org/10.1186/1687-1847-2014-325 Choi, H.H.: A new method for variable structure control system design: A linear matrix inequality approach. Automatica 33, 2089–2092 (1997). MR1486914. https://doi.org/10.1016/S0005-1098(97)00118-0 Ding, X., Li, W.: Stability and bifurcation of numerical discretization Nicholson blowflies equation with delay. Discrete Dyn. Nat. Soc. 2006, 19413 (2006), 12 pp. MR2244293. https://doi.org/10.1155/DDNS/2006/19413 El-Metwally, E., Grove, E.A., Ladas, G., Levins, R., Radin, M.: On the difference equation x n + 1 = α + β x n − 1 e − x n . Nonlinear Anal. 47(7), 4623–4634 (2001). MR1975856. https://doi.org/10.1016/S0362-546X(01)00575-2 Feckan, M., Marynets, K.: Study of differential equations with exponential nonlinearities via the lower and upper solutions’ method. Numer. Anal. Appl. Math. 1(2), 1–7 (2020). https://doi.org/10.36686/Ariviyal.NAAM.2020.01.02.007 Fridman, E., Shaikhet, L.: Stabilization by using artificial delays: an LMI approach. Automatica 81, 429–437 (2017). MR3654628. https://doi.org/10.1016/j.automatica.2017.04.015 Fridman, E., Shaikhet, L.: Delay-dependent LMI conditions for stability of stochastic systems with delay term in the form of Stieltjes integral. In: Proceedings of 57th IEEE Conference on Decision and Control, Fontainebleau, Miami Beach, Florida, USA, pp. 6567–6572 (2018) Fridman, E., Shaikhet, L.: Simple LMIs for stability of stochastic systems with delay term given by Stieltjes integral or with stabilizing delay. Syst. Control Lett. 124, 83–91 (2019). MR3899077. https://doi.org/10.1016/j.sysconle.2018.12.007 Gikhman, I.I., Skorokhod, A.V.: Stochastic Differential Equations. Springer, Berlin, Germany (1972). MR0346904 Gouaisbaut, F., Dambrine, M., Richard, J.P.: Robust control of delay systems: a sliding mode control design via LMI. Syst. Control Lett. 46(4), 219–230 (2002). MR2010239. https://doi.org/10.1016/S0167-6911(01)00199-2 Kara, M.: Investigation of the global dynamics of two exponential-form difference equations systems. Electron. Res. Arch. 31(11), 6697–6724 (2023). MR4657209. https://doi.org/10.3934/era.2023338 Kolmanovskii, V., Shaikhet, L.: Some peculiarities of the general method of Lyapunov functionals construction. Appl. Math. Lett. 15(3), 355–360 (2002). MR1891559. https://doi.org/10.1016/S0893-9659(01)00143-4 Kolmanovskii, V., Shaikhet, L.: About one application of the general method of Lyapunov functionals construction. Int. J. Robust Nonlinear Control 13(9), 805–818 (2003). MR1998313. https://doi.org/10.1002/rnc.846 Li, J., Zhang, B., Li, Y.: Dependence of stability of Nicholson’s blowflies equation with maturation stage on parameters. J. Appl. Anal. Comput. 7(12), 670–680 (2017). MR3602445. https://doi.org/10.11948/2017042 Li, W.T., Fan, Y.H.: Existence and global attractivity of positive periodic solutions for the impulsive delay Nicholson’s blowflies model. J. Comput. Appl. Math. 201(1), 55–68 (2007). MR2293538. https://doi.org/10.1016/j.cam.2006.02.001 Matsumoto, A., Szidarovszky, F.: Delay differential neoclassical growth model. J. Econ. Behav. Organ. 78, 272–289 (2011) Matsumoto, A., Szidarovszky, F.: Asymptotic behavior of a delay differential neoclassical growth model. Sustainability 5, 440–455 (2013) Nakamura, M., Ozawa, T.: Global solutions in the critical Sobolev space for the wave equations with nonlinearity of exponential growth. Math. Z. 231, 479–487 (1999). MR1704989. https://doi.org/10.1007/PL00004737 Nguang, S.K.: Robust H∞ control of a class of nonlinear systems with delayed state and control: an LMI approach. In: Proceedings of 37th IEEE Conference on Decision and Control, Tampa, Florida, USA, pp. 2384–2389 (1998) Nicholson, A.J.: An outline of the dynamics of animal populations. Aust. J. Zool. 2(1), 9–65 (1954) Niculescu, S.I.: H ∞ memory less control with an α-stability constraint for time delays systems: an LMI approach. IEEE Trans. Autom. Control 43(5), 739–743 (1998). MR1618043. https://doi.org/10.1109/9.668850 Papaschinopoulos, G., Ellina, G., Papadopoulos, K.B.: Asymptotic behavior of the positive solutions of an exponential type system of difference equations. Appl. Math. Comput. 245, 181–190 (2014). MR3260707. https://doi.org/10.1016/j.amc.2014.07.074 Papaschinopoulos, G., Fotiades, N., Schinas, C.J.: On a system of difference equations including negative exponential terms. J. Differ. Equ. Appl. 20(5-6), 717–732 (2014). MR3210311. https://doi.org/10.1080/10236198.2013.814647 Papaschinopoulos, G., Radin, M.A., Schinas, C.J.: On the system of two difference equations of exponential form: x n + 1 = a + b x n − 1 e − y n , y n + 1 = c + d y n − 1 e − x n . Math. Comput. Model. 54(11), 2969–2977 (2011). MR2841840. https://doi.org/10.1016/j.mcm.2011.07.019 Papaschinopoulos, G., Radin, M.A., Schinas, C.J.: Study of the asymptotic beha-vior of the solutions of three systems of difference equations of exponential form. Appl. Math. Comput. 218(9), 5310–5318 (2012). MR2870051. https://doi.org/10.1016/j.amc.2011.11.014 Seuret, A., Gouaisbaut, F.: Hierarchy of LMI conditions for the stability analysis of time-delay systems. Syst. Control Lett. 81, 1–7 (2015). MR3353689. https://doi.org/10.1016/j.sysconle.2015.03.007 Shaikhet, L.: Lyapunov Functionals and Stability of Stochastic Difference Equations. Springer Science & Business Media, London, UK (2011). MR3076210. https://doi.org/10.1007/978-3-319-00101-2 Shaikhet, L.: Lyapunov Functionals and Stability of Stochastic Functional Differential Equations. Springer, Berlin, Germany (2013). MR3076210. https://doi.org/10.1007/978-3-319-00101-2 Shaikhet, L.: Stability of equilibrium states for a stochastically perturbed Mosquito popu-lation equation. Dyn. Contin. Discrete Impuls. Syst., Ser. B, Appl. Algorithms 21(2-3), 185–196 (2014). MR3298848 Shaikhet, L.: Stability of equilibrium states for a stochastically perturbed exponential type system of difference equations. J. Comput. Appl. Math. 290, 92–103 (2015). MR3370394. https://doi.org/10.1016/j.cam.2015.05.002 Shaikhet, L.: Stability of equilibriums of stochastically perturbed delay differential neoclassical growth model. Discrete Contin. Dyn. Syst., Ser. B 22(4), 1565–1573 (2017). MR3639178. https://doi.org/10.3934/dcdsb.2017075 Shaikhet, L.: Stability of the zero and positive equilibria of two connected neoclassical growth models under stochastic perturbations. Commun. Nonlinear Sci. Numer. Simul. 65, 86–93 (2019). Pub. online: August 2018. MR3860057. https://doi.org/10.1016/j.cnsns.2018.07.033 Shaikhet, L.: About stability of nonlinear stochastic differential equations with state-dependent delay. Symmetry 14(11), 2307 (2022). https://www.mdpi.com/2073-8994/14/11/2307/pdf Shaikhet, L.: Stability of equilibria of exponential type system of three differential equations under stochastic perturbations. Math. Comput. Simul. 206, 105–117 (2023). MR4517038. https://doi.org/10.1016/j.matcom.2022.11.008 Shaikhet, L.: Stability of the exponential type system of stochastic difference equations. Mathematics 11(18), 3975 (2023). Special Issue “Nonlinear Stochastic Dynamics and Control and Its Applications”. https://www.mdpi.com/2227-7390/11/18/3975 Thai, T.H., Dai, N.A., Anh, P.T.: Global dynamics of some system of second-order difference equations. Electron. Res. Arch. 29(6), 4159–4175 (2021). MR4342296. https://doi.org/10.3934/era.2021077 Wan, H., Zhu, H.: A new model with delay for mosquito population dynamics. Math. Biosci. Eng. 11(6), 1395–1410 (2014). MR3263418. https://doi.org/10.3934/mbe.2014.11.1395