Show Summary Details
More options …

# Open Mathematics

### formerly Central European Journal of Mathematics

Editor-in-Chief: Gianazza, Ugo / Vespri, Vincenzo

IMPACT FACTOR 2018: 0.726
5-year IMPACT FACTOR: 0.869

CiteScore 2018: 0.90

SCImago Journal Rank (SJR) 2018: 0.323
Source Normalized Impact per Paper (SNIP) 2018: 0.821

Mathematical Citation Quotient (MCQ) 2018: 0.34

ICV 2017: 161.82

Open Access
Online
ISSN
2391-5455
See all formats and pricing
More options …
Volume 16, Issue 1

# Boundedness control sets for linear systems on Lie groups

Víctor Ayala
• Corresponding author
• Instituto de Alta Investigación, Universidad de Tarapacá, Sede Esmeralda, Avda. Luis Emilio Recabarren 2477, Iquique, Chile
• Email
• Other articles by this author:
• De Gruyter OnlineGoogle Scholar
/ María Torreblanca Todco
• Departamento Académico de Matemáticas, Universidad Nacional de San Agustín de Arequipa, Calle Santa Catalina, no 117, Arequipa, Peru
• Email
• Other articles by this author:
• De Gruyter OnlineGoogle Scholar
Published Online: 2018-04-18 | DOI: https://doi.org/10.1515/math-2018-0035

## Abstract

Let Σ be a linear system on a connected Lie group G and assume that the reachable set 𝓐 from the identity element eG is open. In this paper, we give an algebraic condition to warrant the boundedness of the existent control set with a nonempty interior containing e. We concentrate the search for the class of decomposable groups which includes any solvable group and also every compact semisimple group.

MSC 2010: 16W25; 93B05; 93C05

## 1 Introduction

Let G be a connected Lie group with Lie algebra 𝔤. In [1] the authors introduce the notion of a linear system Σ on G which is determined by a family of differential equations

$g˙(t)=X(g(t))+∑j=1muj(t)Xj(g(t)),$

where 𝓧 is a linear vector field, Xj ∈ 𝔤 considered as right invariant vector fields and u ∈ 𝓤 ⊂ L(ℝ, Ω ⊂ ℝm) is the class of admissible controls. We deal with Ω as a subset of ℝm with 0 ∈ intΩ. Furthermore, Σ is called restricted if Ω is compact and unrestricted if Ω = ℝm.

We denote by ϕt,u(g) = ϕ(t, g, u) the solution of Σ with control u, initial condition g at time t.

The controllability property of any system is a relevant issue in system theory. It gives you the possibility to connect any two arbitrary states of the manifold through a Σ-solution in positive time. For instance, when G is the Euclidean space ℝn an unrestricted linear system is controllable if and only if it satisfies de Kalman rank condition [1], which is nothing more than the ad-rank condition, see Remark 2.2 in chapter two. However, controllability is rare in the literature, especially for Σ. Assume G is nilpotent and the accessibility set 𝓐 from the identity element eG provided by

$A={x∈G:∃u∈U and t≥0 withx=ϕt,u(e)}$

is an open set. It turns out that

$Σ is controllable on G⇔SpecLy(D)∩R={0}.$

Here, 𝓓 ∈ 𝔤 is a 𝔤-derivation associated to 𝓧 and the Lyapunov spectrum SpecLy(𝓓) consists of the real parts of the 𝓓-eigenvalues.

Recently, the authors in [2] proved that the requirement SpecLy(𝓓) ∩ ℝ = {0} implies controllability for any Lie group with finite semisimple center, that is, for any Lie group which admits a maximal semisimple Lie subgroup with finite center. Certainly, the condition on the Lyapunov spectrum of 𝓓 is very strong. Actually, each 𝓓-eigenvalue must live on the axis iℝ.

For restricted system there exists the notion of control set introduced in [3], basically, a subset 𝓒 where controllability holds on int(𝓒). For a locally controllable system, it is shown in [4] the shape of the control sets with nonempty interior. Under our assumptions, the control set containing the identity element eG reads as

$Ce=cl(A)∩A∗$

where 𝓐 is the reachable set of Σ, i.e., when time in Σ is reversed.

In this paper, we are interested in research on algebraic condition to ensure the boundedness of 𝓒e. We concentrate the study on solvable Lie groups because in this case the space state is firstly decomposable. This means that G can be written as a direct product of the closed subgroups G+, G0 and G with Lie algebras 𝔤+, 𝔤0, and 𝔤 induced by the 𝔤-derivation 𝓓 which determines the drift vector field 𝓧. Secondly, any solvable Lie group has trivially the finite semisimple center property. Hence, we can apply any result about control sets from [5]. In particular, denote by 𝓐G = 𝓐 ∩ G and $\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$ = 𝓐G+. The authors show that for semisimple or nilpotent Lie groups the compactness of 𝓐 G, $\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$ and G0 together is a sufficient condition for the boundedness of 𝓒e. Furthermore, for the class of nilpotent simply connected Lie groups these conditions are also necessary. However, to compute effectively these three sets is a very hard task. Hence, our main aims in this paper are to search for algebraic computable conditions to get the boundedness of 𝓒e. Next, we resume the chapters.

In Section 2, we review some of the standard facts on linear systems. In particular, we summarize without proof the primary relevant material on the dynamic structure, the reachable sets and the existence and uniqueness of control sets with a nonempty interior of Σ. We also mention the 𝓓-decomposition of the Lie algebra 𝔤 and the corresponding Lie groups induced by 𝓧. In Section 3, our main result is stated and proved. A sufficient algebraic condition for the boundedness of 𝓒e is given. In Section 4, we remark some possible extensions. And finally, Section 5 contains a couple of examples in low dimensional Lie groups.

## 2 Preliminaries

In what follows Σ will denote a linear system on a connected Lie group G. In this section, we establish the basic definitions and the main results about the topological and dynamic structure of Σ. In particular, we list some properties of the reachable sets of Σ and we mention the Lie algebra decomposition induces by the drift vector field 𝓧 on 𝔤 = 𝔤+ ⊕ 𝔤0 ⊕ 𝔤 and, its dynamics consequences on the corresponding closed subgroups G+, G0, and G of G.

## 2.1 The dynamic structure of Σ

As we mention, a linear system Σ is furnished by

$g˙(t)=X(g(t))+∑j=1muj(t)Xj(g(t)),u∈U.$

Essentially, its dynamic behavior is determined by two different classes of vector fields. First, the uncontrolled differential equation ġ(t) = 𝓧 (g(t)). Denote by (φt)t∈ℝ the flow of 𝓧. By definition, 𝓧 is an infinitesimal automorphism, which means

${φt:t∈R} is a subgroup of Aut(G),$

where Aut(G) is the Lie group of all G-automorphisms. Associated with 𝓧 there exists a derivation 𝓓 of 𝔤 supplied by

$DY=−[X,Y](e), for all Y∈g.$

The relationship between φt and 𝓓 is given by the formulas, [6],

$(dφt)e=etDandφt(exp⁡Y)=exp⁡(etDY),t∈R,Y∈g.$

On the other hand, the family of vector field $\begin{array}{}{X}^{u}=\sum _{j=1}^{m}{u}_{j}{X}^{j}\end{array}$ depends on m fixed right invariant vector fields Xj ∈ 𝔤 and the family of admissible control u = (u1, …, um) ∈ 𝓤 which has the mission to redirect 𝓧 to reach the desired goal.

## 2.2 Reachable sets

For a state gG, the reachable set from g up to the time t is defined by

$A≤t(g)={h∈G:∃u∈U and τ∈[0,t] withh=ϕτ,u(g)}.$

and 𝓐(g) = ⋃t>0 𝓐t(g) is the reachable set from g. We denote 𝓐(e) by 𝓐.

Next, we collect the main properties of the reachable sets, see [7] and [8].

#### Proposition 2.1

For a linear system Σ on the connected Lie group G it holds

1. 0 ≤ t1t2 implies 𝓐t1 ⊂ 𝓐t2

2. for all gG, 𝓐t(g) = 𝓐tφt(g)

3. for all u ∈ 𝓤, gG and t ≥ 0 it follows ϕt,u(𝓐(g)) ⊂ 𝓐(g)

4. e ∈ int𝓐 if and only if 𝓐 is open

The controllable set to g up to the time t is defined by

$A≤t∗(g)={h∈M:∃u∈U and τ∈[0,t] andϕτ,u(h)=g}.$

The controlled set to g is 𝓐(g) = ⋃t>0𝓐t(g). We denote 𝓐(e) by 𝓐.

#### Remark 2.2

We assume from the start that 𝓐 is an open set and it happens for instance, when the system satisfies the ad-rank condition, i.e.,

$Span{Di(Yj):whereD0=Id,j=1,…,mandi=0,1,…}=g.$

The system is said to be locally accessible at g if int(𝓐t(g)) and int(𝓐t(g)) is nonempty for any t ≥ 0, and controllable from g if 𝓐(g) = G.

## 2.3 𝓓-Decomposable Lie groups

In this section, we look more closely at the Lie algebra decomposition induced by the derivation 𝓓 associated with the drift vector field 𝓧. We address the generalized eigenspaces of 𝓓 provided by

$gα={Y∈g:(D−α)nY=0 for some n≥1}.$

Here, α runs over the spectrum Spec(𝓓). It turns out that [𝔤α,𝔤β] ⊂ 𝔤α+β if α+βSpec(𝓓) and 0 otherwise. Of course, 𝔤 decomposes as

$g=g+⊕g0⊕g−, whereg+=⨁α:Re(α)>0gα,g0=⨁α:Re(α)=0gα and g−=⨁α:Re(α)<0gα.$

It follows that 𝔤+, 𝔤0, 𝔤 are Lie subalgebras and 𝔤+, 𝔤 are nilpotent, see Proposition 3.1 in [6].

Let us denote by G+, G, G0, G+,0, and G−,0 the connected Lie subgroups of G with Lie algebras 𝔤+, 𝔤, 𝔤0, 𝔤+,0 = 𝔤+ ⊕ 𝔤0 and 𝔤−,0 = 𝔤 ⊕ 𝔤0, respectively.

#### Definition 2.3

Let 𝓓 be a 𝔤-derivation. The Lie algebra 𝔤 is said to be 𝓓-decomposable if 𝔤 = 𝔤+ ⊕ 𝔤0 ⊕ 𝔤.

We collect some basic properties of these subgroups, Proposition 2.9 in [7].

#### Proposition 2.4

Let 𝓓 be a 𝔤-derivation. It holds,

1. G+,0 = G+G0 = G0G+ and G−,0 = GG0 = G0G

2. G+G = G+,0G = G−,0G+ = {e}

3. G+,0G−,0 = G0

4. G+,G0,G,G+,0 and G−,0 are closed

5. If G is solvable then G is decomposable

## 2.4 Control sets

A more realistic approach to the controllability property of a system comes from the following notion. A nonempty set 𝓒 ⊂ G is called a control set, [3] if

1. for every gG there exists u ∈ 𝓤 such that ϕ(t, g, u) ⊂ 𝓒, t ≥ 0

2. 𝓒 ⊂ cl(𝓐(g)) for every g ∈ 𝓒

3. 𝓒 is maximal with properties (i) and (ii).

In [4] the authors prove general results about the shape of an existent control set, that we specialize in our particular class of linear systems, as follows:

#### Lemma 2.5

Let 𝓒 be a control set of Σ. If the system is locally accessible at any point of int(𝓒) then for any y ∈ int 𝓒

$C=cl(A(y))∩A∗(y).$

In particular, the system is controllable on int 𝓒.

Instead to study the strong (global) controllability property of Σ we are looking for a weak conditions to obtain regions where controllability holds.

## 3 Main result

In this section, our main results are stated and proved. For that, we apply several results appearing in [5]. From now we assume that G is decomposable. It turns out that 𝓐 and 𝓐 are also decomposable. Denote by 𝓐G = 𝓐 ∩ G and $\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$ = 𝓐G+ then

$A=AG−G+,0,A∗=AG+∗G−,0.$

Furthermore, 𝓐G, $\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$ and G0 are contained in 𝓐 ∩ 𝓐.

We assume that the reachable set 𝓐 is open then the system is locally accessible in a neighborhood of e. From Lemma 2.5, Σ has a control set 𝓒e = cl(𝓐) ∩ 𝓐. On the other hand, by hypothesis 𝔤 is 𝓓-decomposable hence 𝓒e is the only control sets with nonempty interior.

It is clear that

$Ce=cl(A)∩A∗ bounded⇒cl(AG−),cl(AG+∗)andG0 bounded.$

In the sequel, we analyze a kind of converse. Actually, in some special cases, the boundedness of these three sets imply the boundedness of 𝓒e. The following two results were proved in [5].

#### Theorem 3.1

Let us assume that G is semisimple or nilpotent. If cl(𝓐G), cl($\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$) and G0 are compact subsets of G then 𝓒e is bounded.

Recall that a linear transformation L is said to be hyperbolic if L has just eigenvalues with nonzero real parts, in other words

$SpecLy(D)∩R={0}.$

#### Theorem 3.2

Let G be a nilpotent simply connected Lie group. Then,

$Cisbounded⇔cl(AG−),cl(AG+∗)arecompactsandDishyperbolic.$

#### Remark 3.3

The main aim of the paper is to find algebraic conditions to decide wether cl(𝓐G), cl($\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$) are bounded sets. With that and the hyperbolic notion we can ensure the boundedness of the control set 𝓒e.

Let Σ be a linear system on a connected Lie group G. If 𝓓 is a stable matrix the reachable set on G is bounded. More precisely:

#### Proposition 3.4

Let 𝓓 be the derivation associated with 𝓧. If SpecLy(𝓓) ⊂ ℝ then the reachable set 𝓐 = 𝓐 is bounded

#### Proof

Let us denote by ρ the left invariant metric of G, [9]. There exist c > 1 and λ > 0 such that

$ρ(φt(g),e)≤c−1e−λtρ(g,e), for any t≥0,(∗)$

In fact, consider a curve γ : [0, 1] → G with γ(0) = e and γ(1) = g. Thus, φtγ is a curve connecting e to φt(g) and

$ρ(φt(g),e)≤∫01‖(d(φt)γ(s)(γ˙(s))‖ds.$

Now, any G-homomorphism ϕ satisfies the formula

$ϕ∘Lg=Lϕ(g)∘ϕ.$

Subsequently,

$(dϕ)g=(dLϕ(g))e∘(dϕ)e∘(dLg−1)g.$

The homomorphism φt belongs to Aut(G) for any t ∈ ℝ and 𝓓 = d(φt)e. Since the metric ρ is left invariant, we get

$‖d(φt)g‖=‖etD‖.$

By hypothesis 𝓓 is a stable matrix, then (∗) follows.

Take t > 0 such that 𝓐tB(e, 1) the open ball with center e and radius 1. Just observe that for every positive number τ

$φτ(B(e,1))⊂B(e,c−1e−λτ).$

By using the same argument we obtain

$A(n+1)t=Atφt(At) φ2t(At)…φnt(At)⊂B(e,1)B(e,c−1e−λt)B(e,c−1e−2λt)…B(e,c−1e−nλt).$

Now, any gG can be decomposed as

$g=g0g1g2…gn with gi∈B(e,c−1e−itλ),i=0,1,…,n.$

Since the metric is left invariant, the following inequalities are true

$ρ(g,e)=ρ(g0g1g2…gn,e)≤ρ(g0g1g2…gn,g0)+ρ(g0,e)≤ρ(g1g2…gn,e)+ρ(g0,e)≤ρ(g1g2…gn,g1)+ρ(g1,e)+ρ(g0,e)≤ρ(g2…gn,e)+ρ(g1,e)+ρ(g0,e)…≤Σi=0nρ(gi,e)<Σi=0nc−1e−iλt

Hence, there exists a radius R such that

$R=c−1Σi=0∞(e−λt)i>0⇒Ant⊂B(e,R), for any n∈N.$

This ends the proof, actually

$A=∪n∈NAnt⊂B(e,R).$ □

Now, we are able to prove our main result.

#### Theorem 3.5

Let Σ be a linear system on a decomposable connected Lie group G. Assume that 𝔤+,0 is an ideal of 𝔤 then cl(𝓐G) is bounded.

#### Proof

According to our hypothesis the group is decomposable, thus

$G=G−G0G+.$

Since 𝔤+,0 is an ideal the Lie subgroup G+,0 is normal. In particular, the homogeneous space G/G+,0 is a Lie group isomorphic to G. Let us consider the canonical projection π : GG/G+,0. It turns out that π(𝓐) = 𝓐. Furthermore, on G/G+,0 the derivation 𝓓 associated to the drift vector field 𝓧 of Σ has just eigenvalues with negative real parts. In other words, 𝓓 is the corresponding derivation associated with the system Σ in G. In fact, the Lie algebra of G/G+,0 is isomorphic to 𝔤+,0 = 𝔤+ ⊕ 𝔤0 which is isomorphic to 𝔤.

Therefore, Proposition 3.4 implies that the reachable set 𝓐 of Σ is bounded in G. In the sequel, we prove that this condition is enough to show that the reachable set 𝓐 is bounded in G. However, we first need to show that

$πG−:G−→G/G+,0 is a homeomorphism.$

Actually, since any element in G has a unique decomposition in G G0G+ the application is bijective. By the own definition of the quotient topology on G/G+,0 the projection π restricted to G is continuous. Next, we prove that πG is an open map. First, there are neigborhoods VG and WG+,0 of the identity eG such that the product VW is also a neigborhood of e. In particular, πG(V) = π(V) = π(VW) is an open set in G/G+,0. If gG we consider the translations Lg(V) = gV and Lg(W) = gW. Since Lg is a homeomorphism, the proof is done and πG is a homeomorphism.

Once again, the group G is decomposable, thus π(G) = G/G+,0 and it is possible to cover 𝓐 with the projection of a compact subset of 𝓐. In fact, for any compact K containing 𝓐 define the compact set K = (πG)−1(K) ⊂ G such that 𝓐π(K). From that, we obtain

$π(A)=A−⇒A⊂K−G+,0⇒AG−=A∩G−⊂K−.$

Since 𝓐 is bounded, it follows that cl(𝓐 G) is also bounded as we claim. □

#### Corollary 3.6

Let Σ be a linear system on a decomposable connected Lie group G. Assume that 𝔤−,0 is an ideal of 𝔤 then cl($\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$) is bounded.

#### Proof

The proof is completely analogous to that of Theorem 3.5. □

Every nilpotent Lie group as a solvable group is decomposable, see [5].

#### Theorem 3.7

Let Σ be a linear system on a nilpotent simply connected Lie group G. Assume that 𝔤+,0 and 𝔤−,0 are ideals of 𝔤. Then,

1. 𝓓 hyperbolicthe control set 𝓒e = cl(𝓐) ∩ 𝓐 is bounded

2. G = G → 𝓐 = G and 𝓒e = cl(𝓐) is compact

3. G = G+ → 𝓐 = G and 𝓒e = 𝓐 is open

#### Proof

1. We have,

$D hyperbolic ⇔g0={0}⇔g0 is compact⇔G0 is compact.$

The last equivalence depends strongly on the fact that in this particular case the exponential map is a global diffeomorphism. Just observe that in general this is not true. For instance, exp(ℝ) = S1, however, the 1-dimensional sphere is not simply connected. Now, our hypothesis and Theorem 3.5 implies that cl(𝓐G) and cl($\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$) are bounded. Thus, Theorem 3.1 shows that 𝓒e is bounded. On the other hand, if 𝓒e is bounded it follows that G0 ⊂ cl (𝓐) is compact and ending the proof.

2. To prove the second item we observe that under the hypothesis G ⊂ 𝓐. So, 𝓒e = cl(𝓐) is trivially closed and bounded by Theorem 3.1.

3. If G = G+ we get 𝓐 = G. Thus, 𝓒e coincides with the open set 𝓐.

□

#### Remark 3.8

We observe that item third of Theorem 3.7 shows that Σ is controllable from the identity, i.e., for any arbitrary gG there exists a control u and a positive time t such that ϕt,u(e) = g. For other results in the same spirit, we invite the readers to take a look at the following references, [2, 7, 8, 10, 11]. Furthermore, in [12] the author shows that the class of linear control systems is important in a theoretical way. He proves an equivalent theorem which involves a class of nonlinear control systems on general manifolds.

A sufficient condition for the simultaneous boundedness of cl (𝓐G) and cl($\begin{array}{}{\mathcal{A}}_{{G}^{+}}^{\ast }\end{array}$) is to assume that both 𝔤+,0 and 𝔤−,0 are ideals of 𝔤. An equivalent condition is given by the next proposition.

#### Proposition 3.9

Let 𝔤 be a Lie algebra and 𝓓 ∈ 𝔤. It turns out

$g+,0andg−,0areidealsofg⇔[g0,g+]=0and[g+,g−]⊂g0.$

## 4 Extensions

In this paper, we concentrate the study on decomposable Lie groups. However, one might be tempted to try to extend the result to semisimple groups. Let us consider an unrestricted linear system Σ on a connected semisimple Lie group G. In this case, we have two possibilities

1. The compact case

In [10] the authors prove the following result:

#### Theorem 4.1

If G is a connected and compact semisimple Lie group, a linear system Σ is controllable on G if and only if the system is transitive, i.e., satisfies the Lie algebra rank condition, (LARC), provided by

$SpanLA{Di(Yj):whereD0=Id,j=1,…,mandi=0,1,…}=g.$

The LARC condition is weaker than the ad-rank condition. Actually, in the first case you are allowed to compute the Lie brackets [𝓓i1(Yj1),𝓓i2(Yj2)], which is forbidden in the other case. Therefore,

$Ce=GforanytransitivelinearsystemonG.$

• 2

The noncompact case

Here, we just comment that except the case G = G0, the space state cannot be decomposable. In fact, in [5] we show that the set GG0G+G is just an open Bruhat cell which is dense in the group all. In particular, our results can not be extended in this direction.

## 5 Examples

In this section, we give some examples of boundedness and unboundedness control sets on some decomposable Lie groups. But first, we explain how to find the face of the drift 𝓧 when it is induced by an inner derivation.

#### Remark 5.1

A particular class of linear vector fields is easy computed through a 1-parameter of inner G-automorphisms. Take X ∈ 𝔤 a right invariant vector field and consider the solution Xt(g) with initial condition gG. By the right invariance, the solution through the initial condition g is provided by the right translation by g of the solution Xt(e) = expG(tX) through the identity element. In order words

$Xt(g)=expG⁡(tX)g.$

Here, expG:𝔤 → G is the usual exponential map. Hence, X defines by conjugation a 1-parameter group of inner automorphism as follows

$φt(g)=Xt(e)gX−t(e),g∈G, and φt∈Aut(G) for any t∈R.$

Therefore, it is possible to compute the linear vector field as

$X(g)=(ddt)t=0φt(g).$

The associated derivation 𝓓 : 𝔤 → 𝔤 is 𝓓(Y) = −[X, Y], Y ∈ 𝔤.

Recall that any derivation on a semisimple Lie group is inner. This property has interest for us in the compact case. On the other hand, in [13] we built an algorithm which provides an effective means to compute the Lie algebra 𝔤 that we use in this section.

#### Example 5.2

Consider the solvable affine group

${G=(xy01):x>0andy∈R}$

with Lie algebra 𝔤 = Span {X, Y} and [X, Y] = Y. An easy computation shows that 𝔤 is given just by inner derivation with the shape

$∂g={D=(00ab):a,b∈R}.$

From Remark 5.1 the linear vector fiel 𝓧 associated to 𝓓 is given by

$X(x,y)=(0a(x−1)+by00).$

Let Σ be the transitive linear system on G defined by

$g˙(t)=X(g(t))+u(t)X(g(t)),u∈U$

where, 𝓓 = ad(Y) comes from a = −1 and b = 0. Since ad(Y)X = −Y then Span {X,𝓓X} = 𝔤. So, Σ satisfies the ad-rank condition, 𝓐 is open and of course, Σ satisfies also LARC. Moreover, G is solvable thus the control set 𝓒e is the only one with nonempty interior. It turns out that,

$g+=g−=0⇒g0=g.$

Thus,

$G+,0=G⊂A and G−,0=G⊂A∗⇒Ce=G.$

To conclude, the system is controllable from the identity. This fact is completely concordant with Theorem 3 in [11]. Actually, it is shown there that a transitive system in a canonical form, like Σ, is controllable if and only if b = 0.

#### Example 5.3

Let 𝔤 = ℝX + ℝY + ℝZ be the Lie algebra of the connected and simply connected Heisenberg Lie group G

$G={g=(1xz01y001):(x,y,z)∈R3}$

of dimension three. The generators of 𝔤 are provided by

$X=∂∂x,Y=∂∂y+x∂∂z and Z=∂∂z.$

The only one non-vanishing Lie bracket is [X, Y] = Z. Any derivation 𝓓 is represented by a 6 real parameters matrix in the basis {X, Y,Z} as follow

$∂g={(ab0cd0efa+d):a,b,c,d,e,f∈R}.$

Consider the linear system Σ with derivation 𝓓 determined by its coefficients a = d = −1, b = 1, c = −1, e = f = 0 and the control vectors X and Z,

$g˙(t)=X(g(t))+u1(t)X(g(t))+u2(t)Z(g(t)),u∈U, with Ω=[−1,1].$

We have, SpecLy(𝓓) = {−1, −2}. So, 𝔤−,0 = 𝔤 = 𝔤 and 𝔤+,0 = 0 are both ideals of 𝔤. On the other hand,

$Span{X,Z,D(X)=X−Y}=g.$

Since 𝓓 is a hyperbolic derivation, Theorem 3.7 shows that the existent control set 𝓒e is bounded.

#### Example 5.4

On the rotational group S0(3, ℝ) with Lie algebra 𝔰𝔬 (3, ℝ) the skew-symmetric real matrix of order three

$g=Span{X,Y,Z}$

consider the system

$g˙(t)=X(g(t))+u1(t)X(g(t))+u2(t)Y(g(t)),u∈U,withΩ=R,$

where 𝓧 = ad(X). Since Σ satisfies LARC, the control set is bounded and coincides with the group. The system is controllable from the identity.

#### Example 5.5

Take the linear system Σ on the Heisenberg group G like in Example 2, but with different dynamic behavior

$g˙(t)=X(g(t))+u1(t)(X−Y)(g(t))+u2(t)(X+Y+Z)(g(t)),u∈U,withΩ=[−1,1]$

where the derivation 𝓓 is furnished by a = 1, d = −1 and b = c = e = f = 0. Hence, 𝓓(XY) = X + Y. Thus,

$Span{X−Y,Z,X+Y}=g$

an 𝓐 is an open set.

If we restrict Σ to the plane ℝ2 = Span {X, Y} we get a classical linear system on the vector space Σ2

$(x˙y˙)=(100−1)(xy)+(11)u:u∈U with Ω=[−1,1]$

which satisfies Lemma 2.5. Moreover,

$A=R×(−1,1) and A∗=(−1,1)×R.$

Thus, the control set 𝓒e restricted to the plane is bounded and reads

$(Ce)R2=(−1,1)×[−1,1], see [4].$

However, 𝓒e can not be bounded. Despite the fact that 𝔤+,0 = Span {X, Z} and 𝔤−,0 = Span {Y, Z} are ideals, the derivation 𝓓 = diag(1, −1, 0) is just hyperbolic on the plane not on G. Actually,

$Ce=(−1,1)×[−1,1]×(1,1,1)R.$

#### Example 5.6

Let us consider the nilpotent Lie group G with Lie algebra

$g=RX1+RX2+RX3+RX4,andtherules[X4,X2]=X4,[X3,X2]=X4+X2,[X1,X2]=X3and[ X1,X3]=X4.$

Let Σ be a linear system with an arbitrary derivation 𝓓 ∈ 𝔤 such that the reachable set 𝓐 of Σ is open. Hence, the control set 𝓒e is unbounded. In fact, a straightforward computation shows that the Lie algebra of 𝔤-derivations is five dimensional and reads as

$∂g={(a−a000000bca+b0deb+c+d2a+b):a,b,c,d,e∈R}.$

Since the underlying topological space of G is the connected and simply connected manifold ℝ4, Theorem 3.5 applies. However, 0 ∈ SpecLy(𝓓) for any 𝓓 ∈ 𝔤. Thus, no hyperbolic derivation exits, ending the proof.

## Acknowledgement

Proyecto de Investigación Básica e Investigación Aplicada, n 24, 2017, Concytec-Fondecyt, UNSA, Perú.

## References

• [1]

Ayala, V. and Tirao, J., Linear control systems on Lie groups and Controllability, Eds. G. Ferreyra et al., Amer. Math. Soc., Providence, RI, Vol. 64, 99. 47-64,1999. Google Scholar

• [2]

Ayala, V. and Da Silva, A. J., Controllability of linear systems on Lie groups with finite semisimple center, SIAM Journal on Control and Optimization, Vol. 55, n° 2, pp. 1332-1343, 2017.

• [3]

Kliemann, W., Qualitative Theory Nichtlinearer Stochastischer Systeme. Dissertation, Universitat Bremen, 1980. Google Scholar

• [4]

Colonius, F. and Kliemann, The Dynamics of Control, Systems & Control: Foundations & Applications. Birkäuser Boston, Inc., Boston, MA, 2000. Google Scholar

• [5]

Ayala, V., Da Silva A. J. and Zsigmond, Control sets of linear systems on Lie groups. Non Differential Equations and Applications, Vol 24, Issue 1, Article number 8, pp. 1-15, 2017. Google Scholar

• [6]

San Martin, L. A. B., Algebras de Lie, Second Edition, Editora Unicamp, (2014). Google Scholar

• [7]

Da Silva, A. J, Controllability of linear systems on solvable Lie groups. SIAM Journal on Control and Optimization 54 No. 1 (2016), pp. 372-390.

• [8]

Jouan, Ph., Controllability of Linear Systems on Lie group, Journal of Dynamics and Control Systems, 17 (2011) 591-616.

• [9]

Onishchik, A. L. and E. B. Vinberg, Lie Groups and Lie Algebras III - Structure of Lie Groups and Lie Algebras, Berlin: Springer (1990). Google Scholar

• [10]

Ayala, V. and L.A.B. San Martin, Controllability properties of a class of control systems on Lie groups, Lecture Notes in Control and Information Sciences 258 (2001), pp. 83-92. Google Scholar

• [11]

Jouan, Ph. and Dath M., Controllability of Linear Systems on low dimensional Nilpotent and Solvable Lie groups, Journal of Dynamics and Control Systems, April 2016, Volume 22, Issue 2, pp. 207-225.

• [12]

Jouan, Ph. Equivalence of Control Systems with Linear Systems on Lie Groups and Homogeneous Spaces, ESAIM: Control Optimization and Calculus of Variationes, 16 (2010) 956-973. Google Scholar

• [13]

Ayala, V., Kizil and Azevedo Tribuzy, I., On an algorithm for finding derivations of Lie algebras, Proyecciones Journal of Mathematics, Vol. 31, n°1, pp. 81-90, 2012.

## About the article

Accepted: 2018-01-25

Published Online: 2018-04-18

Citation Information: Open Mathematics, Volume 16, Issue 1, Pages 370–379, ISSN (Online) 2391-5455,

Export Citation