• Keine Ergebnisse gefunden

On the Maximal Output Admissible Set for a Class of Bilinear Discrete- time Systems

N/A
N/A
Protected

Academic year: 2022

Aktie "On the Maximal Output Admissible Set for a Class of Bilinear Discrete- time Systems"

Copied!
18
0
0

Wird geladen.... (Jetzt Volltext ansehen)

Volltext

(1)

http://dx.doi.org/10.1007/s12555-020-0486-6 http://www.springer.com/12555

On the Maximal Output Admissible Set for a Class of Bilinear Discrete- time Systems

Youssef Benfatah, Amine El Bhih, Mostafa Rachik, and Abdessamad Tridane*

Abstract:Given a discrete-time controlled bilinear systems with initial statex0and output functionyi, we investi- gate the maximal output setΘ(Ω) ={x0∈Rn,yi∈Ω,∀i≥0}whereΩis a given constraint set and is a subset ofRp. Using some stability hypothesis, we show thatΘ(Ω)can be determined via a finite number of inequations.

Also, we give an algorithmic process to generate the setΘ(Ω). To illustrate our theoretical approach, we present some examples and numerical simulations. Moreover, to demonstrate the effectiveness of our approach in real-life problems, we provide an application to the SI epidemic model and the SIR model.

Keywords:Asymptotic stability, bilinear systems, constraint set, discrete-time systems, output admissible set.

1. INTRODUCTION

Output admissible sets arise in many essential practi- cal applications such as stability analysis and control of constrained systems (see [1–6]). This concept has been considerably investigated (see [7,8] and the references therein), but not in the case of the bilinear systems. Con- sequently, its applicability is limited.

Bilinear systems were introduced into control theory in the 1960s. They were a distinct type of nonlinear systems, in which nonlinear terms are developed by way of multi- plication of control vector and state vector. These type of systems has attracted a large community of researchers in almost half a century (see [9]). Such systems’ significance lies in the fact that many necessary engineering processes [10] can be modeled through bilinear systems [11].

As we know, a nonlinear discrete-time finite- dimensional system can be represented by the following dynamic equation:

xi+1= f(xi,ui), (1)

wherexi∈Rnandui∈Rq, fori≥0, are respectively the state and the control.

If we assume that f is linear with respect to the state or the control, we obtain, respectively.

xi+1=h1(ui)xi+f1(ui), (2)

and

xi+1=h2(xi)ui+f2(xi), (3) we obtain an another form of bilinear systems given as follows:

xi+1=Axi+Bx˜ iui+Bui, (4) where for each i≥0, A is a n×n matrix,B is a n×q matrix. ˜Bxiuiis a bilinear form in thexi,uivariables that can be expressed as

Bx˜ iui=

q

j=1

uijBjxi, (5)

whereui={uij}qj=1andBjis an×nmatrix.

The use of bilinear systems has two significant advan- tages: First, they provide better modeling of a nonlinear phenomenon than a linear system [12]). Second, many real-life problems can be model using bilinear systems.

1.1. Related work

The maximal output admissible (MOA) sets are an important concept for analyzing controlled systems with constraints. The MOA set has been well studied particu- larly for linear systems with state and control constraints [13,14]. This concept provides an understanding of the analysis of constrained control systems. Besides, it is widely used in control system design methods [15].

Manuscript received July 13, 2020; revised November 27, 2020 and January 6, 2021; accepted January, 31, 2021. Recommended by Associate Editor Ohmin Kwon under the direction of Editor Jessie (Ju H.) Park. The authors would like to thank the reviewer for his time to help improve this paper. Research reported in this paper was supported by the Moroccan Systems Theory Network.

Youssef Benfatah, Amine El Bhih, and Mostafa Rachik are with the Laboratory of Analysis Modeling and Simulation,Department of Mathematics and Computer Science, Faculty of Sciences Ben M’Sik, Hassan II University, Casablanca, Sidi Othman, BP 7955 (e-mails:

{youssef.benfatah, elbhihamine}@gmail.com, m-rachik@yahoo.fr). Abdessamad Tridane is with the Department of Mathematical Sci- ences,United Arab Emirates University, Al Ain P.O. Box 15551, UAE (e-mail: a-tridane@uaeu.ac.ae).

* Corresponding author.

ICROS, KIEE and Springer 2021c

(2)

The MOA set for a class of nonlinear systems has re- cently been studied by Rachiket al. in [16]. Later on, Hi- rata and Ohta [17] considered a special class of nonlinear systems, the so-called polynomial systems, and discussed the finite determinability of the MOA set.

Some other resources that provide information about the study of the concept MOA sets are given in the ref- erences [18–24]. In [18], Ossareh considered a Lyapunov- stable periodic system and gave the development of the MOA set theory. He also investigated some of its geomet- ric and algebraic properties. In [19], Yamamoto presented MOA’s computational procedure for a trajectory tracking control of biped robots.

Various algorithms have been added in the literature for determining the maximal state constraint sets [13,25]. In [25] the authors considered autonomous linear discrete- time systems subject to linear constraints, and proposed an effective procedure to determine the maximal set of admissible initial states. Other models from population dynamics and optimal control problems can be found in [26,27].

1.2. Problem statement

This work aims to present a new approach to studying the MOA set for a class of bilinear discrete-time systems.

To the best of our knowledge, the MOA sets for this type of system have not been investigated before. More specif- ically, we characterize the initial states of a controlled bi- linear system whose resulting trajectory verify a point- wise constraint. Moreover, our approach is applicable to epidemiological, sociological, and physiological models.

In these examples, the MOA set represents the set of ini- tial data that guarantee for these systems do not violate certain constraints.

This paper considers discrete-time controlled bilinear systems. More precisely, the system has the following form:

xi+1=Axi+

q

j=1

uijBjxi+Bvi, i≥0, x0∈Rn,

(6)

the corresponding output is

yi=Cxi, i∈N, (7)

whereA∈ L(Rn,Rn)is the matrix of the state,R,is the set of real numbers,L(Rn,Rn),is the set of real matrices of ordern×n,B∈ L(Rm,Rn)is the matrix of the input, C∈ L(Rn,Rp)is the matrix of the output,xi∈Rn is the state variable,Bj∈ L(Rn,Rn)for all j∈ {1, 2, ...,q},N, is the set of nonnegative integers,n,m, p andq are the nonnegative integer anduij,vi are the feedback controls defined by

uij=Kjyi and vi=Lyi (8)

withL∈ L(Rp,Rm)andKj∈ L(Rp,R),j∈ {1, ...,q}are matrices with single row, and are given by

Kj=h

k1j k2j k3j . . . kp−1j kpj

i

. (9)

Our purpose right here is to determine all vectors x0

(the initial states) such that the corresponding signal out- put(yi)iverify the conditionyi∈Ω,∀i∈N, whereΩ⊂Rp is a given constraint set. The set of all such vectors is the maximal output admissible setΘ(Ω).

For technical consideration we restrict our own selves in the current work to the study of Θλ(Ω), the maximal outputλ-admissible set, which defined by

Θλ(Ω) =Θ(Ω)∩B(0,λ), (10) whereB(0,λ)is the ball of center 0 and radiusλ>0, this means we study the set of all vectors x0∈Rn such that kx0k ≤λ andyi∈Ω,∀ i∈N, whereλ is any positive number.

Moreover, we give some useful properties of Θλ(Ω), and we investigate its determination by way of a finite number of functional inequalities.

The rest of this paper is structured as follows: In Section 2, we present some preliminary results. In Section 3, we characterize the maximal outputλ-admissible set. More- over, in Section 4 we give some sufficient conditions to ensure the finite determination of the maximal outputλ- admissible set. In the following section, we propose an algorithm for determining the output admissibility index and, consequently the set Θλ(Ω). Various numerical ex- amples to illustrate our results are given in Section 6. An application to a controlled SI epidemic model and SIR model is given, to demonstrate the effectiveness of our ap- proach in Section 7. The last section includes conclusion.

In the next section, we will give some necessary math- ematical preliminaries that will be used in this paper.

2. PRELIMINARY RESULTS

In this section we consider the discrete-time controlled bilinear systems described by (6)-(7).

We replaceuijandviby its values in system (6) we ob- tain

xi+1= (A+BLC)xi+

q

j=1

KjCxiBjxi, i≥0, x0∈Rn.

(11) Notations: In what follows we will denote the matrix A+BLCby ˜A,KjCby ˜Kj and{i,i+1, ...,k},i≤k, the finite subset ofNbyσik.

Remark 1: In caseq=1 the system (11) becomes (xi+1=Ax˜ i+δxiB1xi, i≥0

x0∈Rn

(12)

(3)

whereδ=K1C∈ L(Rn,R).

Hence and by taking into account the previous nota- tions, we consider in this paper the class of discrete sys- tems defined by

xi+1=Ax˜ i+

q

j=1

jxiBjxi, i≥0, x0∈Rn,

(13)

increased by the output

yi=Cxi, ∀i≥0, (14)

and we are interested in the theoretical and numerical characterization of the set

Θλ(Ω) ={x0∈B(0,λ)∩Rn|yi∈Ω, ∀i∈N}. (15) Indeed, letψ be the function defined as follows:

ψ: Rn→Rn x 7→ψ(x) =

q

j=1

jxBjx. (16) LetKbe the matrix defined by

K=

k11 k12 · · · k1p k21 k22 · · · k2p ... ... . .. ... kq1 kq2 · · · kqp

. (17)

Definition 1: LetΩ⊂Rpandx0∈Rn. The initial state x0is said to beΩ−output admissible if

yi∈Ω, ∀i∈N.

The set of all such initial states is defined by Θ(Ω) ={x0∈Rn|yi∈Ω, ∀i∈N}. The general solution of system (6) is given by

xii(x0), ∀i∈N, (18) where the functionΦis defined by

Φ: Rn→Rn

x 7→Φ(x) =Ax˜ +ψ(x), and

Φi=

i−times

z }| { Φ◦Φ◦...◦Φ.

Using (18) the setΘ(Ω)could be rewritten as follows:

Θ(Ω) =

x0∈Rn|CΦi(x0)∈Ω, ∀i∈N .

In this paper we restrict our-selves to the study of the set Θλ(Ω)defined by

Θλ(Ω) =Θ(Ω)∩B(0,λ)

={x0∈B(0,λ)∩Rn|CΦi(x0)∈Ω,∀i∈N}.

(19) The restriction to the setΘλ(Ω)will not reduce the value of the work and this for the following reason. Having an initial state x0∈Rn, we would like to know if x0 is Ω−output admissible or not. In order to give an answer to such a question, we first identify the set Θλ(Ω) = Θ(Ω)∩B(0,λ)whereλ is a real that verifies kx0k ≤λ and we check ifx0∈Θλ(Ω)or not.

Remark 2: 1) The general solution of system (6) is given by

xi=

i−1

k=0

A˜+

q

j=1

ukjBj

!

x0,∀i≥1. (20) 2) ∀x,y∈Rn,

kψ(x)−ψ(y)k

≤ ( q

j=1

k(K˜j)>kkBjk(kxk+kyk) )

× kx−yk, (21) where>denotes the transpose.

3) Forx0∈B(0,λ), we have

kxik ≤Fi(λ), ∀i∈N (22) whereFis defined by

F(t) =αt+CKt2, ∀t∈R, (23) with

α=kAk˜ and CK=

q

j=1

k(K˜j)>kkBjk.

Now, we give some conditions which are sufficient to ensure, for convenient initial statesx0,the asymptotic sta- bility of system (6). Our main result in this direction is the following.

Proposition 1: Suppose the following hypothesis to hold.

1) There exists τ≥1 and θ ∈]0,1[ such that kAk˜ k ≤ τ θk,∀k∈N( the matrixLcan be chosen such that the hypothesis (1) holds ),

2) θ+CKλ τ2<1,where CK=

q

j=1

k(K˜j)>kkBjk. (24)

(4)

Then the system (6) is asymptoticaly stable in the ball of center 0 and radiusλ, i.e.,

i→∞limkxik=lim

i→∞i(x0)k=0,∀x0∈B(0,λ). (25) Proof: Letx0∈B(0,λ). Then the general solution of system (6) can be written as

xi=A˜ix0+

i−1

k=0

i−k−1ψ(xk), ∀i≥1 (26)

=⇒ kxik ≤ kA˜ikkx0k+

i−1 k=0

kA˜i−k−1kkψ(xk)k

=⇒ kxik ≤τ θikx0k+

i−1

k=0

τ θi−k−1kψ(xk)k.

On the other hand, we have kψ(xk)k=k

q

j=1

jxkBjxkk (27)

q

j=1

k(K˜j)>kkxkkkBjkkxkk (28)

=

q

j=1

k(K˜j)>kkBjk

!

kxkk2 (29)

=CKkxkk2, (30) we have used Cauchy-Schwarz inequality. This leads to

kxik ≤τ θikx0k +CKτ θi−1

i−1 k=0

θ−kkxkk2, ∀i≥1.

Letzi=xi

λ. Then kzik ≤τ θikz0k

+λCKτ θi−1

i−1

k=0

θ−kkzkkkzkk, ∀i≥1. (31) Sincekz0k ≤τ, we prove that

kzik ≤τ, ∀i∈N. (32) Indeed, if we assume that

kzik ≤τ, ∀i∈σ0N, (33) then using (31) it follows

kzik ≤τ θikz0k +λCKτ2θi−1

i−1

k=0

θ−kkzkk, ∀i∈σ1N+1. (34) Take

yi−ikzik and Γi=α+β

i−1

k=0

yk, (35)

where

α=τkz0k and β=λCKτ2θ−1. (36) From (34) we conclude

yi≤Γi, ∀i∈σ1N+1. (37) We haveyi=Γi+1−Γi

β . Then Γi+1−Γi

β ≤Γi, ∀i∈σ1N+1, which implies

yi≤Γi≤(1+β)i−1Γ1, ∀i∈σ1N+1. (38) If one replacesyiandΓ1by their values, we deduce from (38) that∀i∈σ1N+1,

kzik ≤ θ+CKλ τ2i

τ+CKλ τ2θ−1

1+CKλ τ2θ−1 kz0k. (39) Because,θ+CKλ τ2<1 andkz0k ≤1 it follows

kzik ≤τ, ∀i∈σ1N+1, and

kzik ≤τ, ∀i∈N. (40) Using (39) for alli≥0, we deduce that

i→∞limkzik=0 sinceθ+CKλ τ2≤1.

Therefore

i→∞limkΦi(x0)k=0. (41)

This completes the proof.

The following result assumes that 0∈Ω (

Ωdenoted the interieur ofΩ), this assumption is satisfied in any reason- able application and has nice consequences (see [13] ). In practice, the setΩhas the form

Ω={y∈Rp| f1(y)≤0, f2(y)≤0, . . . , fs(y)≤0}, (42) wheres∈Nand fi,i∈σ1sare continuous functions such that fi(0)≤0,∀i∈σ1s, such a sets have many importance in a practical view.

Remark 3: The condition taken above fi(0)≤0,∀i∈ σ1simplies 0∈Θλ(Ω)and then the setΘλ(Ω)nonempty.

Indeed, the setΘλ(Ω)is given by Θλ(Ω)

={x0∈B(0,λ)∩Rn|CΦi(x0)∈ Ω, ∀i∈N}

(5)

={x0∈B(0,λ)|fj(CΦi(x0))≤0, ∀j∈σ1s,∀i∈N}, and we have

fj(CΦi(0)) =fj(0)≤0, ∀j∈σ1s,∀i∈N, (43) sinceΦ(0) =0.

Hence 0∈Θλ(Ω)because 0∈B(0,λ).

Imposing special conditions on ˜A, ˜Kj,Bj, j∈ {1, ..., q} and Ω which imposes corresponding conditions on Θλ(Ω).

Proposition 2: 1) IfΩis closed, then the setΘλ(Ω)is also closed.

2) If we assume the following assumptions to hold:

(a) there existsτ≥1 andθ∈]0,1[such thatkAk˜ k≤ τ θk,∀k∈N,

(b) θ+CKλ τ2≤1,CK=

q

j=1

k(K˜j)>kkBjk, (c) 0Rp∈Ω,

then 0Rn

Θ\λ(Ω).

Proof: 1) Leti∈N. Then we define the functions Ψi:B(0,λ)∩Rn→Rp, x07→CΦi(x0). (44) Using these functions, the setΘλ(Ω)can be written as

Θλ(Ω) =\

i∈N

Ψ−1i (Ω). (45) SinceΩis closed andΨi,i≥0 are continuous func- tions (sinceψ(x) =

q

j=1

KjxBjxis continuous),Ψ−1i (Ω) is closed and thenΘλ(Ω)is closed.

2) By hypothesis (a), (b), and (c), and from Proposition 1 we have

∀z0∈B(0,λ), lim

i→∞i(z0)k=0. (46)

This leads to

∀z0∈B(0,λ), ∀ε>0, ∃i0∈N: kΦi(z0)k ≤ε, ∀i≥i0. i.e.,

∀z0∈B(0,λ), ∀ε>0, ∃i0∈N:

Φi(z0)∈B(0,ε), ∀i≥i0. On the other hand

0∈Ω =⇒ ∃η>0 : B(0,η)⊂Ω. (47) Forε=kCkη , we have

∀z0∈B(0,λ), ∃i0∈N

such that

i(z0)∈B(0,η)⊂Ω,∀i≥i0. (48) It remains to show that

i(z0)∈Ω, ∀i∈σ0i0−1. (49) Φi0−1continuous in 0 implies

∃ηi0−1>0, ∀z0∈B(0,ηi0−1),

i0−1(z0)−Φi0−1(0)k ≤ η kCk

=⇒ ∃ηi0−1>0, ∀z0∈B(0,ηi0−1), CΦi0−1(z0)∈Ω.

Φi0−2continuous in 0 implies

∃ηi0−2>0, ∀z0∈B(0,ηi0−2), CΦi0−2(z0)∈Ω by a similar way, using the continuity ofΦj, j∈σ1i0−3 in 0, we obtain

∀j∈σ1i0−3,

∃ηj>0,∀z0∈B(0,ηj),CΦj(z0)∈Ω. (50) We have also

∀z0∈B

0, η kCk

,Cz0∈Ω. (51)

If we takeς=inf(ηi0−1, ...,η1,kCkη ), we get

∀z0∈B(0,ς),CΦi(z0)∈Ω, ∀i∈σ0i0−1. (52) If we choose this timeζ =inf(λ,ς), it follows

∀z0∈B(0,ζ),CΦi(z0)∈Ω, ∀i∈N. (53) ThusB(0,ζ)⊂Θλ(Ω), and consequently 0∈

Θ\λ(Ω).

This completes the proof.

3. CHARACTERIZATION OF THE MAXIMAL OUTPUTλ-ADMISSIBLE SET

It is more difficult to characterize the setΘλ(Ω)defined in (19), for that reason we define for eachk∈Nthe set

Θλk(Ω) ={x0∈B(0,λ)∩Rn:CΦi(x0)∈Ω,∀i∈σ0k}, (54) and we introduce the following definition.

Definition 2: The setΘλ(Ω)is said to be finitely de- termined if there exists an integerk0such thatΘλ(Ω) = Θλk

0(Ω).

Remark 4: 1) Forr,s∈Nsuch thatr≤s, we have Θλ(Ω)⊂Θλs(Ω)⊂Θλr(Ω). (55)

(6)

2) IfΘλ(Ω)is finitely determined, andk0the smallestk such thatΘλk(Ω) =Θλk+1(Ω), then we have

Θλ(Ω) =Θλk0(Ω) =Θλk(Ω),∀k≥k0. (56) Conditions that demonstrate finite determinability, are examined in the following proposition.

Proposition 3: 1) IfΘλ(Ω)is finitely determined then there existsk0∈Nsuch thatΘλk0(Ω) =Θλk0+1(Ω).

2) IfΦ(B(0,λ))⊂B(0,λ)and Θλk0(Ω) =Θλk0+1(Ω)for somek0∈NthenΘλ(Ω)is finitely determined.

Proof: 1) Assume that the setΘλ(Ω)is finitely deter- mined. Then by Definition 2 it follows

∃k0∈Nsuch thatΘλ(Ω) =Θλk0(Ω). (57) Using Remark 4 we have Θλk0+1(Ω)⊂Θλk0(Ω)since k0+1≥k0. On the other hand,

Θλ(Ω)⊂Θλk0+1(Ω). (58) We deduce from (57) and (58) that

Θλk0(Ω)⊂Θλk0+1(Ω). (59) Therefore

Θλk0(Ω) =Θλk0+1(Ω) for some k0∈N. (60) 2) Letx0∈Θλk

0(Ω), thenx0∈Θλk

0+1(Ω).

This leads

x0∈B(0,λ), CΦi(x0)∈Ω, ∀i∈σ0k0+1. (61) it follows

i(Φ(x0))∈Ω, ∀i∈σ0k0. (62) ThusΦ(x0)∈Θλk0(Ω)since

Φ(x0)∈Φ(B(0,λ))⊂B(0,λ). (63) By iteration

x0∈Θλk0(Ω) =⇒Φj(x0)∈Θλk0(Ω),∀j∈N

=⇒CΦi Φj(x0)

∈Ω, ∀i∈σ0k0

=⇒CΦi(x0)∈Ω, ∀i∈N

=⇒x0∈Θλ(Ω) sincex0∈B(0,λ). Hence

Θλk0(Ω)⊂Θλ(Ω). (64) Using Remark 4, we conclude

Θλk0(Ω) =Θλ(Ω)for somek0∈N, (65)

which completes the proof.

4. SUFFICIENT CONDITIONS FOR FINITE DETERMINATION OFΘλ(Ω)

The following two theorems are the main results that give simple conditions to guarantee the finite determina- tion of the setΘλ(Ω).

Theorem 1: Assume the following hypothesis to hold (i) There existsτ≥1 andθ∈]0,1[such thatkAk˜ k

τ θk,∀k∈N,

(ii) θ+CKλ τ2≤1,CK=

q

j=1

k(K˜j)>kkBjk, (iii) 0∈Ω,

(iv) Φ(B(0,λ))⊂B(0,λ).

Then,Θλ(Ω)is finitely determined.

Theorem 2: If we assume that

(i) kΦ(z)k ≤νkzk,∀z∈Rnandν∈]0,1[. (ii) 0∈Ω.

Then,Θλ(Ω)is finitely determined.

5. ALGORITHMIC DETERMINATION

In this section we notek0the smallest integer such that Θλ(Ω) =Θλk

0(Ω), (66)

we callk0the admissibility index. We shall provide a use- ful algorithm (see [13]) for identifying this admissibil- ity index k0 and therefore the characterization of the set Θλ(Ω).

Remark 5: 1) Algorithm 1 producek0 andΘλ(Ω)if and only ifΘλ(Ω)is finitely determined.

2) If Algorithm 1 does not converge then the setΘλ(Ω) is not finitely determined.

Algorithm 1 is not practical due to the fact it does not describe how the testΘλk(Ω) =Θλk+1(Ω)is implemented.

To deal with this difficulty we will consider the following approach.

LetΩbe defined as

Ω={y∈Rp/f1(y)≤0,f2(y)≤0, . . . ,fs(y)≤0}. (67)

Algorithm 1:Determination ofk0.

Require: λ>0,n,p,q∈N, ˜A,C,ψ,Ω⊂Rp k←0

if Θλk(Ω) =Θλk+1(Ω) then letk0←k and stop else

letk←k+1 and continue (return to previous test ) end if

(7)

Algorithm 2Determination ofk0

Require: λ >0,n,p,q,s∈N,A,C,˜ ψ,fi,i=1, ...,s k←0

Repeat fori=1,...,sdo

MaximizeJi(x) =fi(CΦk+1(x)) Subjet to the constraints (fjl(x)

≤0, x∈B(0,λ),

∀j∈ {1, . . . ,s}, ∀l∈ {0, . . . ,k}.

end for

Ji←max{Ji(x)}

if Ji≤0,∀i=1,2, . . . ,s then k0←k

break else

k←k+1 end if

Thus, for allk∈N, the setΘλk(Ω)can be written as Θλk(Ω) =

x∈B(0,λ)|fj(CΦi(x))≤0, j=1, . . . ,s, i=0, . . . ,k . (68) On the other hand,

Θλk+1(Ω) =

x∈Θλk(Ω,K)|fjk+1(x)

≤0, j=1, . . . ,s . (69) Now, sinceΘλk+1(Ω)⊂Θλk(Ω)for allk∈N, we deduce

Θλk+1(Ω) =Θλk(Ω,K)

⇐⇒Θλk(Ω)⊂Θλk+1(Ω)

⇐⇒ ∀x∈Θλk(Ω),fjk+1(x)

≤0,∀j∈ {1, . . . ,s}

⇐⇒ sup

x∈B(0,λ),fi(CΦl(x))≤0

∀i∈{1,...,s},∀l∈{0,...,k}

fjk+1(x)

≤0,∀j∈ {1, . . . ,s}.

Therefore, the test Θλk(Ω) =Θλk+1(Ω) leads to a set of mathematical programming problems. We will suggest another version of Algorithm 1, this new algorithm is given by Algorithm 2.

Remark 6: If the matrix ˜Ais Lyapunov stable then the supremum in Algorithm 2 is defined for allx∈Rn. Indeed, using the notation defined in Proposition 1 it follows

kCΦ(x)k=

CAx˜ +C

q

j=1

jxBx

(70)

≤ CAx˜

+kCkCKkxk2. (71) On the other hand, the assumption of Lyapunov stability of A˜implies

CAx˜

≤γkxkfor some positiveγ. This leads kCΦ(x)k ≤(γ+kCkCKkxk)kxk. (72)

Letk∈Nand define the functionFby

F(α) = (γ+kCkCKα)α, ∀α∈R. (73) Then

k+1(x)

≤Fk+1(kxk), (74) and

k+1(x)∈B(0,Fk+1(kxk)), (75) using the compactness ofBand the continuity of the func- tions fi, the sequence fik+1(x)

is bounded from above.

Remark 7: 1) This algorithm can not be effectively used if there is not a feasible way to find global op- tima. WhenΩ is a polyhedron (i.e., the functions fj

are affine for all j), the difficulty becomes less.

2) Assumptions of our two previous results (Theorem 1 and 2 of Section 4) are sufficient but not necessary. If these assumptions are not established, then the conver- gence of Algorithm 2 is not guaranteed.

To illustrate our results, we should give several exam- ples, which will be given in the upcoming section.

5.1. Algorithm 2 steps In casek=0 we have fori=1

maxf1(CΦ(x)) s.c

(f1(Cx)≤0, . . . , fs(Cx)≤0,

|x1| ≤λ, . . . , |xn| ≤λ, fori=2

maxf2(CΦ(x)) s.c

(f1(Cx)≤0, . . . , fs(Cx)≤0,

|x1| ≤λ, . . . , |xn| ≤λ, we continue like this untili=s

maxfs(CΦ(x)) s.c

(f1(Cx)≤0, . . . , fs(Cx)≤0,

|x1| ≤λ, . . . , |xn| ≤λ.

If max fi(CΦ(x))≤0,∀i∈ {1, . . . ,s}then we stop and k0=0 else we continue.

In casek=1 we have fori=1

maxf1(CΦ2(x)) s.c





f1(Cx)≤0, . . . , fs(Cx)≤0, f1(CΦ(x))≤0, . . . , fs(CΦ(x))≤0,

|x1| ≤λ, . . . , |xn| ≤λ,

(8)

fori=2

maxf2(CΦ2(x)) s.c





f1(Cx)≤0, . . . , fs(Cx)≤0, f1(CΦ(x))≤0, . . . , fs(CΦ(x))≤0,

|x1| ≤λ, . . . , |xn| ≤λ, we continue untili=s

maxfs(CΦ2(x)) s.c





f1(Cx)≤0, . . . , fs(Cx)≤0, f1(CΦ(x))≤0, . . . , fs(CΦ(x))≤0,

|x1| ≤λ, . . . , |xn| ≤λ.

If maxfi(CΦ2(x))≤0,∀i∈ {1,. . .,s}then we stop and k0=1 else we continue.

For example in stepk=kwe have fori∈ {1,. . .,s}

maxfi(CΦk+1(x))

s.c

















f1(Cx)≤0, . . . , fs(Cx)≤0, f1(CΦ(x))≤0, . . . , fs(CΦ(x))≤0,

...

f1k(x)

≤0, . . . , fsk(x)

≤0,

|x1| ≤λ, . . . , |xn| ≤λ.

If max fi(CΦk+1(x))≤0, ∀i∈ {1, . . . ,s}then we stop andk0=k else we continue. We can use a predefined output function to be called at each iteration.

Remark 8: Let us suppose that the cost of function of maximization is Costmax, the cost of repeat loop is

CostRepeat and the cost associated to the computation of

i,i∈ {0, 1,. . .,CostRepeat}isCostf ct. Then the compu- tational complexity of Algorithm 2 is given by

Costalgo= (s×Costmax+2s)×CostRepeat+Costf ct. (76) 6. NUMERICAL EXAMPLES

To illustrate our results, we present some examples in two and three dimensional cases (n=2 andn=3). For this reason, we construct the matricesKj,j∈σ1q (in the case whenp=1) by the following way

Kj= 1

Nb× kBjk × kCk ×q×λ×τ2,

where the choice of the numberNb>1 depend on the value ofθ which satisfies the first hypothesis of Propo- sition 1. This choice of Kj guarantee the verification of the second hypothesis of Proposition 1. The matrix ˜A is chosen such thatkAk˜ <1. Even ifkAk>1, this choice is possible by selecting some matrixLsince ˜A=A+BLC.

Using these matrices, we guarantee the asymptotic stabil- ity of the considered systems. AssumptionΦ(B(0,λ))⊂ B(0,λ)is also taken into consideration. In all examples the dotted region will denote the setΘλ(Ω). For Example 4, we have k0 =∞, which means that Algorithm 2 does not converge. Various selections of matrices determining the systems are considered.

In Examples 1-7, we constructΩas follows:

Ω={y∈Rp|fi(y)≤0, i=1, . . . ,2p},

where fi:Rp−→Rare defined for allx= (x1,. . .,xp)∈ Rpby

(f2m−1(x) =xm−ε, m∈ {1,2, . . . ,p}, f2m(x) =−xm−ε, m∈ {1,2, . . . ,p},ε>0, clearly 0∈Ω, fi,i∈σ1sare continuous functions such that

fi(0)≤0,∀i∈σ1s.

Example 1: Letλ,qandndefined by λ =1

2, q=1, n=2.

Let the matrix ˜A,B1,Cand ˜K1defined by A˜=

1

6 1 1 6

8 0

, B1=

1 0 0 0

, K˜1= 641 0

, C= 1 0

.

Then the functions Φ(x),CΦ(x)andCΦ2(x)are defined by

Φ x

y

= 1

64x2+16x+16y

1 8x

,

CΦ x

y

= 1 64x2+1

6x+1 6y, CΦ2

x y

= 1 36x+ 1

36y+ 1 64

1 64x2+1

6x+1 6y

2

+ 1 384x2.

Using Algorithm 2 withε=0.2 we getk0=0 Θλ(Ω) =

x y

∈R2| |x| ≤ 1 2,|y| ≤ 1

2,|x| ≤ε

,

and using the same algorithm withε=0.1 we getk0=1

Θλ(Ω) =





 x

y

∈R2|

|x| ≤ 1

2, |y| ≤1 2,

1 64x2+1

6x+1 6y

≤ε





 .

(9)

Fig. 1.The setΘλ(Ω)corresponding to Example 1 (ε= 0.2).

Fig. 2.The setΘλ(Ω)corresponding to Example 1 (ε= 0.1).

Example 2: Letλ,q,nandεbe defined as λ =1, q=1, n=2, ε=0.3.

Let the matrix ˜A,C, ˜K1andB1defined as A˜=

1 6

1 1 4 2

1 8

, B1=

1 0 2 0

, K˜1= 18 0

, C= 1 0

. Then the functionsΦ

x y

,CΦ x

y

andCΦ2 x

y

are de- fined by

Φ x

y

= 1

8x2+16x+14y

1

4x2+12x+18y

,

CΦ x

y

=1 8x2+1

6x+1 4y,

Fig. 3.The setΘλ(Ω)corresponding to Example 2.

2 x

y

= 1

512x4+ 1

192x3+ 1

128x2y+ 25 288x2 + 1

96xy+11 72x+ 1

128y2+ 7 96y.

Using Algorithm 2 we obtain k0 =1, and then the set Θλ(Ω)

Θλ(Ω) =



 x

y

∈R2|

|x| ≤1,|y| ≤1,|x| ≤ε,

1 8x2+1

6x+1 4y

≤ε



 .

Example 3: Letλ,q,nandεbe defined as λ =2, q=2, n=2 andε=0.1.

Let the matrix ˜A,C, ˜K1, ˜K2,B1andB2be defined as A˜=

1 212

2 3

1 3

, B1=

1 2 0

1 2 0

, B2=

−1 0

−1 0

, K˜1= 321 0

, K˜2= 641 0

, C= 1 0

.

Then Φ

x y

= 1

2x−12y

2 3x+13y

, CΦ

x y

=1 2x−1

2y, Φ2

x y

=

121x−125y

5 9x−29y

, CΦ2

x y

=−1 12x−5

12y, Φ3

x y

=

2372x−727y

7 54x−1954y

, CΦ3

x y

=−23 72x−7

72y.

Using Algorithm 2 we obtain k0 =2, and then the set Θλ(Ω)

Θλ(Ω) =











 x

y

∈R2|

|x| ≤λ,|y| ≤λ,|x| ≤ε,

1 2x−1

2y

≤ε,

−1 12x− 5

12y

≤ε











 .

(10)

Fig. 4.The setΘλ(Ω)corresponding to Example 3.

Example 4: Letλ,q,nandεdefined by λ =2, q=2, n=2, ε=0.3.

Let the matrix ˜A,B1,B2,C,K˜1and ˜K2defined as A˜=

1 2

1 1 2 3

2 3

, B1=

1 2 0

1 2 0

,B2=

−1 0

−1 0

, C= 1 0

, K˜1= 321 0

,K˜2= 641 0 .

Then Φ

x y

= 1

2x+12y

1 3x+23y

,CΦ

x y

=1 2x+1

2y, CΦ2

x y

= 5 12x+ 7

12y,CΦ3 x

y

=29 72x+43

72y, CΦ4

x y

=173 432x+259

432y, CΦ5

x y

=1037

2592x+1555 2592, CΦ6

x y

= 6221

15552x+ 9331 15552y, CΦ7

x y

=37325

93312x+55987 93312y.

Using Algorithm 2 we getk0=∞.

Example 5: Letλ,q,nandεdefined by λ =2, q=2, n=2, ε=0.1.

Let the matrix ˜A,B1,B2,C, ˜K1and ˜K2defined as A˜=

1

8 1 1 4 3

1 24

,B1=

3 0 1 0

, B2= 2 0

3 0

, C= 1 0

,K˜1= 641 0

,K˜2= 1281 0 .

Fig. 5.The setΘλ(Ω)corresponding to Example 5.

Then, the functionsΦ x

y

,CΦ x

y

andCΦ2 x

y

are de- fined by

Φ x

y

= 1

16x2+18x+14y

5

128x2+13x+241y

,

CΦ x

y

= 1 16x2+1

8x+1 4y, CΦ2

x y

= 1

4096x4+ 1

1024x3+ 1

512x2y+ 19 1024x2 + 1

256xy+ 19 192x+ 1

256y2+ 1 24y.

Using Algorithm 2 we getk0=1

Θλ(Ω) =



 x

y

∈R2|

|x| ≤2,|y| ≤2,|x| ≤ε,

1 16x2+1

8x+1 4y

≤ε



 .

Example 6: Let ˜A,B1, ˜K1,C,q,εandλ defined as C= 1 0 0

,λ=2,q=1, ε=0.05, B1=

 1 0 4 2 3 0 1 0 1

,A˜=

1 6

1 4

1 1 5 3

1 8

1 7

0 101 0

,K˜1= 401 0 0 .

Then, we have Φ

 x y z

=

1 6

1 4

1 1 5 3

1 8

1 7

0 101 0

 x y z

+ 1 40x

 1 0 4 2 3 0 1 0 1

 x y z

=

1

6x+14y+15z+101xz+401x2

1

3x+18y+17z+403xy+201x2

1

40x2+401zx+101y

,

(11)

 x y z

=1 6x+1

4y+1 5z+ 1

10xz+ 1 40x2,

Φ2

 x y z

=

1

9x+2400223y+42029z+

1 40

1

6x+14y+15z+101xz+401x22

+

1 10

1

40x2+401zx+101y

×

1

6x+14y+15z+101xz+401x2 +

3

160xy+60013xz+60013x2

7

72x+6720761y+84071z+

1 20

1

6x+14y+15z+101xz+401x22

+

3 40

1

6x+14y+15z+101xz+401x2

×

1

3x+18y+17z+403xy+201x2 +

3

320xy+84031xz+336061 x2

1

30x+801y+701z+

1 40

1

6x+14y+15z+101xz+401x22 +

1 40

1

40x2+401zx+101y

×

1

6x+14y+15z+101xz+401x2 +

3

400xy+2001 x2

 ,

2

 x y z

=1

9x+ 223 2400y+ 29

420z + 1

40 1

6x+1 4y+1

5z+ 1 10xz+ 1

40x2 2

+ 1 10

1 40x2+ 1

40zx+ 1 10y

× 1

6x+1 4y+1

5z+ 1 10xz+ 1

40x2

+ 3

160xy+ 13

600xz+ 13 600x2. Using Algorithm 2 we obtaink0=2, and then

Θλ(Ω) =









































 x y z

∈R3|

|x| ≤λ,|y| ≤λ,

|y| ≤λ,|x| ≤ε,

1

6x+14y+15z +101xz+401x2

≤ε,

1

9x+2400223y+42029z +401

1

6x+14y+15z +101xz+401x2

2

+101 1

40x2+401zx +101y

× 1

6x+14y+15z +101xz+401x2

+1603 xy+60013xz+60013x2

≤ε









































 .

Example 7: To show the efficiency of the stability we take the following example.

Letλ,q,n,εandτdefined as

λ =1, q=1, n=2, ε=0.1, τ=1.

Fig. 6.The setΘλ(Ω)corresponding to Example 8.

Fig. 7.The setΘλ(Ω)corresponding to Example 9.

LetB1, ˜A, ˜K1,C,CKandθdefined by B1=

0 0 0 1

,A˜=

0.1+γ 0

0 0

,C= 1 −1

, K˜1= 101 0

,CK= 1

10, θ= A˜

. Now, we use Algorithm 2 we obtain

γ θ+CKλ τ2 k0

0 0.2 0

0.5 0.7 1

0.6 0.8 1

0.9 1.1 3

1 1.2 ∞.

6.1. Comment

In Examples 1-7, we have determined the index of ad- missibility k0 using the predefined function fmincon in

(12)

Matlab [28], which allow us to solve the maximization problems in Algorithm 2. In Figs. 1-5, we have plotted all the constraints forming the setsΘλ(Ω),in order to vi- sualize such sets of Examples 1, 2, 3, and 5.

Example 8: In this example we takeΩin the form Ω=

y∈R| f(y) =y2+y−2≤0 .

It is easy to see that 0∈Ω, f is continuous function such that f(0)≤ 0 and that Ω is closed since Ω =

f−1(]−∞,0]).

Now, we use the results of Example 3 and Algorithm 2 it follows

Maximize f(CΦ(x,y)) = (1 2x−1

2y)2+(1 2x−1

2y)−2 s.c





f C x y

!!

=x2+x−2≤0,

|x| ≤2, |y| ≤2,

the solution of this problem gives 1.7500.

Maximize f(CΦ2(x,y)) =

−1 12x− 5

12y 2

+

−1 12x−5

12y

−2

s.c

















f C x y

!!

=x2+x−2≤0,

f CΦ x y

!!

= (12x−12y)2+(12x−12y)−2≤0,

|x| ≤2, |y| ≤2,

the solution of this problem gives:−1.2000e−06. Therefore, the value of index of admissiblity isk0=1.

Hence Θλ(Ω)

= x

y

∈R2| |x| ≤2, |y| ≤2,x2+x−2≤0,

x 22y2

+ x22y

−2≤0

= x

y

∈R2| |x| ≤2, |y| ≤2, x2+x−2≤0

∩ x

y

∈R2|x 2−y

2 2

+x 2−y

2

−2≤0

= [−2,1]×[−2,2]

∩ x

y

∈R2|x−2≤y≤x+4

.

Example 9: In this example the setΩis represented by Ω=

y∈R| f1(y) =y3+y−3≤0, f2(y) =y2+2y−4≤0

.

It is easy to see thatfi(0)≤0,∀i∈ {1, 2}and the functions fi,i=1, 2 are continuous.

Using results of Example 5 and Algorithm 2 we ob- taink0=0, and then the maximal outputλ-admissible set Θλ(Ω)is given by

Θλ(Ω) =

 x

y

∈R2|

|x| ≤2, |y| ≤2, x3+x−3≤0, x2+2x−4≤0

= [−2,2]×[−2,2]∩h

−2,√ 5−1i

×[−2,2]

∩[−2,r]×[−2,2]

= [−2,r]×[−2,2],

withr= 3 q3

2181√ 3√

247+ 3 q1

18

√3√

247+32.

6.2. Comment

In Examples 8 and 9, we have modified the setΩwhich affects the mathematical programming problems that arise in Algorithm 2, we have plotted again the setsΘλ(Ω)in Figs. 6 and 7.

In order to obtain the numerical computational (elapsed time in seconds (s)) of Examples 1, 2, 3, 5, and 6 using Algorithm 2 the predefined output functionfmincon(see [28]) is applied with different values for the coefficientλ, ε,k0,nto solve the maximization problems which arise in such algorithm. The obtained results are presented in Table 1.

Remark 9: In case p≥2 we have to reconstruct the matrices Kj such that the condition θ+λCKτ2 <1 of Proposition 1 is verified.

Table 1.Parameters values and elapsed time of Examples 1-3, 5, and 6.

Examples A˜ λ s k0 ε Time (s)

Examples 1

1 6

1 6 1

8 0

!

1

2 2 0 0.2 1.660970

Examples 1

1 6

1 6 1

8 0

!

1

2 2 1 0.1 3.1711

Examples 2

1 6

1 4 1 2

1 8

!

1 2 1 0.3 3.2184

Examples 3

1 212

2 3

1 3

!

2 2 2 0.1 3.0743

Examples 5

1 8

1 4 1 3

1 24

!

2 2 1 0.1 3.5962

Examples 6

1 6

1 4

1 5 1 3

1 8

1 7

0 101 0

 2 2 1 0.05 8.5662

Referenzen

ÄHNLICHE DOKUMENTE

Theo- rem 1 implies upper semicontinuity of 2, also in this topology, but under t h e condition of strong asymptotic stability of the equilibrium set KO with

The author studies the problem of exact local reachability of infinite dimensional nonlinear control systems.. The main result shows that the exact local

The description of this &#34;solution tube&#34; is important for solving problems of guaranteed estimation of the dynamics of uncertain systems as well as for the solution of

Tan, K.C., Optimal Control of Linear Econometric Systems with Linear Equality Constraints on the Control Variables, International Economic Review, Vol. 20,

This report deals with the following questions: which dynamic models and which advanced methods of identification theory are used or could be used in urban

To solve the problem, we have to determine when the activities of the N set will be or can be performed if the supply of resources (vector R(t)) is given for all t within

For this purpose we will base our construction on a dynamic programming technique using the optimal value function of a discounted optimal control problem.. Note that this

The value function and the corresponding optimal control values for each point can also be used to ”verify” the assumptions of Theorem 3.1, (viii) numerically: if there exists a