The regular free-endpoint linear quadratic problem with
indefinite cost
Citation for published version (APA):
Trentelman, H. L. (1989). The regular free-endpoint linear quadratic problem with indefinite cost. SIAM Journal on Control and Optimization, 27(1), 27-42. https://doi.org/10.1137/0327003
DOI:
10.1137/0327003
Document status and date: Published: 01/01/1989 Document Version:
Publisher’s PDF, also known as Version of Record (includes final page, issue and volume numbers) Please check the document version of this publication:
• A submitted manuscript is the version of the article upon submission and before peer-review. There can be important differences between the submitted version and the official published version of record. People interested in the research are advised to contact the author for the final version of the publication, or visit the DOI to the publisher's website.
• The final author version and the galley proof are versions of the publication after peer review.
• The final published version features the final layout of the paper including the volume, issue and page numbers.
Link to publication
General rights
Copyright and moral rights for the publications made accessible in the public portal are retained by the authors and/or other copyright owners and it is a condition of accessing publications that users recognise and abide by the legal requirements associated with these rights. • Users may download and print one copy of any publication from the public portal for the purpose of private study or research. • You may not further distribute the material or use it for any profit-making activity or commercial gain
• You may freely distribute the URL identifying the publication in the public portal.
If the publication is distributed under the terms of Article 25fa of the Dutch Copyright Act, indicated by the “Taverne” license above, please follow below link for the End User Agreement:
www.tue.nl/taverne
Take down policy
If you believe that this document breaches copyright please contact us at:
openaccess@tue.nl
providing details and we will investigate your claim.
THE REGULAR FREE-ENDPOINT
LINEARQUADRATIC
PROBLEMWITH
INDEFINITE COST*
HARRY L. TRENTELMAN
Abstract. This paper studies anopenprobleminthe context of linear quadratic optimalcontrol,the free-endpoint regular linear quadraticproblemwith indefinitecostfunctional.Itisshown that the optimal Cost for thisproblem is given bya particular solution of the algebraic Riccati equation. This solution is characterized intermsof the geometryonthe latticeof all real symmetric solutions of the algebraic Riccati equationas developed by Willems [1EEE Trans. Automat. Control, 16 (1971), pp.621-634] and Coppel [BullAustral. Math. Soc., 10(1974),pp.377-401]. Anecessary and sufficient condition is established for the existence of optimal controls. This condition is stated intermsofasubspace inclusion involving the extremal solutions of thealgebraicRiccati equation. The optimal controlsareshowntobegenerated bya feedback control law. Finally, the results obtained are compared with "classical" results onthe linear quadraticregulator problem.
Keywords, linearquadraticoptimalcontrol,indefinitecostfunctional,free-endpointproblem
AMS(MOS)subjectclassifications. 93C05,93C35,93C60
1. Introduction.
In
this paper we are concerned with regular, infinite-horizon linearquadratic optimal control problems in whichthe cost functional is the integral of anindefinite
quadratic form.In
mostofthe existingliterature ontheregularlinear quadratic(LQ)
problem,it is explicitly assumedthatthe quadratic form inthe cost functional, apartfrom being positive definite in the control variable alone, is positive semidefinite in the control andstatevariablessimultaneously.In fact,
under this semidefinitenessassumptiontheLQ
problem has become quitestandard and istreatedin manybasictextbooks inthe fieldof systems and control[1], [2], [9], [21].
Often a distinction is made between two versions ofthe problem, thefixed-endpoint version and thefree-endpointversion.In
thefixed-endpointversionit
is necessaryto minimize the costfunctionalunder the constraint that the optimal state trajectory should converge to zero as time tends to infinity, while in the free-endpoint version it is.only necessary to minimize the cost functional.For
the case that the quadratic form in the cost functional is positive semidefinite both versions of the regularLQ
problem are well-understood and completelysatisfactory solutions oftheseproblems
are available.Surprisingly,
however,
forthe mostgeneralformulationof theregularLQ
problem, thatis,the case that thequadraticform inthe
costfunctional isindefinite,asatisfactory treatment does not yet exist.In
this case we can again distinguish between the fixed-endpoint version and the free-endpoint version. While for the fixed-endpoint versionacompletesolutionhasbeen described in17]
(see
also14]),
the free-endpoint versionhas onlybeen considered in17]
underaveryrestrictiveassumption. Thuswe seethat,
uptonow,thefree-endpoint regularLQ
problemwithindefinite cost functional has been an open problem. In the present paper we shall fill up this gap and present afairly complete solution to thisproblem.It is well known
[12], [19]
that for the free-endpoint regularLQ
problem with positive semidefinite cost functional, the optimalcost is givenby the smallestpositive semidefinite real symmetric solution ofthe algebraicRiccati equation.Wewill see that this statement is no longervalid in general ifthe cost functional is theintegral ofanReceivedbythe editors November2,1987;acceptedfor publication(inrevisedform)March28,1988.
?Departmentof Mathematics andComputer Science, Eindhoven University ofTechnology,5600MB Eindhoven, the Netherlands.
indefinitequadraticform.
It
willbeshown, however,
thatin this case also the optimal cost is given by a solution ofthe algebraic Riccati equation. Thisparticular solution willbecharacterized in termsof the geometryonthesetofall realsymmetricsolutions ofthe algebraic Riccati equation as described in[17]
and[4].
Another well-known fact is
that,
for the free-endpoint regularLQ
problemwith positive semidefinite costfunctional, theexistenceof optimal controlsis never an issue" under the assumption that the underlying system iscontrollable,
for thisproblem
unique optimal controls always existforall initial conditions. This is in contrast with the fixed-endpoint
LQ
problem,wherethe existence of optimal controls forall initial conditionsdepends onthe"gap"
of the algebraicRiccatiequation(i.e.,
thedifference between thelargest
and smallest solutions of the Riccatiequation). In
this paperwe will see thatalso,
for the free-endpoint regularLQ
problem withindefinite
cost functional, optimal controls no longer needtoexistforall initialconditions!Moreover,
we will establish a necessary and sufficient condition in terms of the"gap"
of the algebraicRiccatiequation for the existence of optimalcontrolsforall initial conditions.We
willshow that forthe particular casethatthe cost functional ispositivesemidefinite this condition is always satisfied, thus explaining the fact that in this special case optimalcontrols alwaysexist. Finally, we willshow that alsoin the indefinite casethe optimal controls for the free-endpoint regularLQ
problem, ifthey exist, aregiven by a feedback control law.The outline ofthis paper is as follows.
In
the remainder ofthis section we will introduce mostofthe notational conventions that willbe used.In
2 wegive formula-tions ofboth the free-endpointandfixed-endpoint regularLQ
problemsthat we shall be dealingwith.In
3 we will briefly recall themostimportantfacts thatweneed on thegeometryofthesetofall realsymmetricsolutions tothe algebraicRiccatiequation as developed in[17]
and[4].
In
4 we will state the solution to the fixed endpoint regularLQ
problem with indefinite cost as established in[17].
Also,
we will stateits(incomplete)
counterpart, the solution to the free-endpoint regularLQ
problem with positive semidefinite cost functional. Then in 5 we will state and prove our maintheorem,
a solution to the free-endpoint regularLQ
problem.In
orderto establish a proof ofthis theorem we will state and prove a series ofsmaller lemmas.In
6 we will show how the "classical" results on the free-endpoint regularLQ
problem with positivesemidefinite costfunctionalcanbe reobtained as aspecialcase
ofourgeneral
solution.
We
will close this paperin 7 with some concluding remarks.We
use the following notational conventions.For
a given nxn matrixA
its set of eigenvalues will be denoted byo-(A).
IfV
is asubspace ofR"
andA
is an nxn matrixthenAI
v
will denote the restrictionofA
toV. V
will be called A-invariantifA
VcV. In
this casetr(AI
v)
willdenotethesetof eigenvalues ofA V
andtr(Al"
/ v)
will denote the set of eigenvalues ofthe mapping induced byA
in the factor spaceffn/V (see [21]).
We will denote subsets of C byC-:={sClRes=0},
CO:
{s
CIRe
s=0},
and C/:=
{s
CIRe
s> 0}.
Given a real monic polynomial p there is auniquefactorizationp p_ "Po’P/into real monicpolynomialswithp_,Po,andp/having all roots in
C-,
C,
and C/,
respectively. IfA
is a real nxn matrix and ifp denotes its characteristic polynomial then we defineX-(A):=
kerp_(A),
X(A):
ker
po(A),
andX/(A):=
kerp/(A).
Thesesubspaces are A-invariant and therestriction ofA
toX-(A)(X(A),
X/(A))
has characteristicpolynomialP-(Po, P/).
A
subset Cg of C will be called symmetric if a+
biC
g:>
a biCg.
If Cg isgiven then we define
Ch
:=C\Cg.
IfA
is areal nxnmatrix and if p is its characteristic polynomialthen,
again,p can be factored uniquely into p=pg’pb, wherepg
and pbXg(A)
:= kerpg(A)
andXh(A):=
kerph(A).
Again thesesubspacesare A-invariant and therestrictionofA
toXg(A)(Xb(A))
has characteristicpolynomialpg(pb).
In fact,
the subspaceXg(A)(X(A))
is equalto the linearspan ofall generalized eigenvectors ofA
corresponding to its eigenvalues inCg(Cb).
Alternatively,Xg(A)(Xh(A))
is equal tothe largest A-invariant subspaceV of"
suchthattr(A[V)c
C(C).
If,
in additiontoA,
arealp xn matrixC
is given, thenwe denote(ker
CIA):=
f
kC
ai-’,
i=1
the unobservable subspace of
(C,
A)
[21,
3.2].
Given asymmetric subsetC
of C we denoteXdet
:=(ker
CIA)
Xb(A),
the undetectable subspace of
(C,
A)
with respect toC
g. The pair(C, A)
is calleddetectable
with respect toC
ifA
is(Cg_)
stable on the unobservable subspace of(C, A),
i.e., if(ker
CIA)c
Xg(A)
(see
[21,
3.6]).
It
is easyto see that(C,
A)
isdetectable if andonlyifXdet
0.Also,
(C, A)
is detectable if and only iffor
all hC
we have ker(A-hi) f3ker C
=0(see
[15]).
In
order to be rigorous on the interpretation ofthe cost functionals thatwill be consideredin
thispaper, we will nowexplainwhat we meanby thestatement that the limit ofa function exists ine.
Let
R
:=U
{-oo,
+oo}.
Givenf:
we say thatlim,_.f(t)
exists if it is equal to a real number in the usual sense.We
say thatlim,_.f(t)
=-o(+o)
ifforallre
there existsTR
suchthat >T
impliesf(t)
<-r(>-r).
Then wesay that lim,_f( t)
exists inR
if itexists, isequalto-c,
orisequal to+.
If
M
is a real n n matrix andV
is a subspace of",
then we defineM-V:=
{x
"lMx
V}.
IfV
is asubspace of[" thenV
-
denotesits orthogonal complement with respectto the standard Euclidean inner product.Finally, we will denote by
L2,oc(R+)
the spae ofall measurable vector-valued functions on+
that aresquare integrableover allfinite intervals inI1+.
L2(
+)
denotes thespaceof
all measurable vector-valued functions on+
that are square integrable over+.
Finally,L(
+)
denotes thespace
of all measurable vector-valued functions on+
that are essentially bounded on+.
Here,
+
:={t
lt
=>
0}.
2. TheregularLQ-problem. Consider the finite-dimensional linear time-invariant system
(2.1)
Ax
+
Bu,
x(O)
Xo.Here,
x and u are assumed to taketheirvalues inRn
andR",
respectively.A
andB
are real nxn and nxrn matrices, respectively.
It
will be astanding assumption that(A, B)
is controllable.We
shall consideroptimizationproblems of the type(2.2)
infto(x,
u)
dt,
where
to(x,u)
is a real quadratic form on InXl defined byto(x,u):=
u
TRu
+
2urSx
+
xTQx.
Here
R, S,
andQ
are assumed to be real matrices such thatthe form to.
For
a given control function uL2,oc(E+),
letX(Xo, u)
denote the state trajectory of(2.1)
and if T_>-0letIo
Jr(xo, u):=
to(X(Xo,
u)(
t),
u( t))
dt.We
now explain how(2.2)
should be interpreted. First, we specify two classes of control functions with respectto whichthe infimization in(2.2)
shouldbeperformed. DefineU(xo)
:={u
L2,.oc(+)]
lirn
Jr(xo, u)
exists ine},
U,.(Xo)
:={u
U(xo)llirn
X(Xo,
u)(t)=
0}.
Note that,
due to the assumption that(A,
B)
iscontrollable,
we haveU(xo)#
andU,.(Xo)
forall Xo".
For
uU(xo)
we define(2.3)
J(xo, u):=
limJ-(Xo, u).
Tx
We
note thatJ(xo, u)
Re.
Now,
define(2.4a)
V.(Xo)
:=inf{J(xo,
u)[u
U(xo)},
(2.4b)
V+(xo)
:=inf{J(xo,
u)[ ]u
U(xo)},
the optimal cost forthe free-endpoint problem and fixed-endpoint problem, respec-tively.
By
the factthat(A,
B)
is controllable we have thatV.;:.(Xo),
V+(xo)
for all x0 [". Following
[17],
we want toexclude the situationthat for certain initial conditions Xo the values(2.4a)
or(2.4b)
become equal to-.
It
can be shown thata necessarycondition forV.;:.
(Xo)>-
andV+(xo)>-
for all Xo to hold is that R_->,0(see [17],
[12]).
In this paper a standing assumption will be that R>
0. Under. this assumption theLQ
problems definedby(2.4)
are called regular.The fixed-endpoint regular
LQ
problem,definedby(2.4b),
wascompletelyresolved in[17] (see
also[14]).
There,
asatisfactorycharacterizationwas givenfor the optimal cost, necessaryand sufficient conditionsweregivenfortheexistenceofoptimalcontrolsfor, all initial conditions, and these optimal controls were given in the form of a state-feedbackcontrol law. The problems ofhowto calculate the optimal costfor.the free-endpointregular
LQ
problem(2.4a),
to state necessary and sufficient conditions for the existence ofoptimal controls, and to calculate theseoptimal controls haveup to now been open.In
this paperwe will considertheseproblems..3.
Geometry
of the algebraic Rieeati equation.A
central role in infinite horizon regularlinear quadratic control problems is played bythe algebraic Riccati equation(ARE)
(3.1)
ArK/KA/Q-(KB/Sr)R-(BK/S)-O.
Let F
denote the setofall real symmetricsolutions oftheARE. It
was shown in[17]
that ifF
is nonemptythen it contains aunique element K/andaunique element K-such that
r(A-
BR-’(BTK
++
S))m
C-Id C,
r(A-
BR-I(BTK
+
S))
C+L3 C.
Moreover,
K+ andK-have the additional property thattheyarethe extremal solutions of theARE
in the sense that ifK
F
then K--<_K K+.
LetA:=K/-K
-.
Denote
A-BR-(BTK++S)
andA-BR-(BTK-+S)
byA
/and
A-,
respectively. If KF
defineAK
:=A-
BR-(BTK +
S). Note
thatX+(A+)
0 andX-(A-)=0.
Let
denote the set of all A--invariant subspaces contained inX+(A-).
The followingbasictheorem is a generalization by Coppel[4]
of a theorem that was originally proven byWillems in[17] (see
also[16],
[10]).
THEOREM 3.1.
Let
(A,
B)
becontrollable,
andassumethatF
is nonempty.If
Visan A--invariant subspace
of X+(A
-)
(that
is,if
V)
thenn=
VA-
V+/-. There exists a bijection y"-->
F
defined
byy(V)
:=K-Pv
+
K+(I-
Pv),
where
Pv
is the projector onto V along A-
V-.
If
K
y(V)
thenX+(AK
V,
X(AK
kerA,
X-(AK
X-(A+)
CI
A-
V-.
Among
other things,the result above statesthatthere existsaone-to-one correspon-dence between the set of all real symmetric solutions oftheARE
and the set of all A--invariant subspaces ofX+(A-).
Following[3],
ifK
3’(V)
then we will say that the solutionK
is supported by the subspaceV.
The next theorem from[4]
states that this one-to-one correspondence in fact respects the partial orderings on the setsF
and
.
THEOREM 3.2. Let
(A, B)
be controllable and assume thatF
is nonempty.Let
KI
and
K2
be solutions to theARE
supported byV
and V2, respectively. ThenK1
<-K2 if
and onlyif
V2
cV.
From the above it
follows,
for example,thatK-
is supported byX+(A
-)
and thatK
+ is supported by0.4. Classical results. In the present section we briefly summarize the solution of the fixed-endpoint regular
LQ
problem with indefinite cost functional as outlined in[17].
Subsequently, we will state the well-known result on the free-endpoint regularLQ
problem with positivesemidefinite
cost functional. Finally, we will discuss some ofthe difficulties that can be expected in trying to generalize the latter result to the case that the semidefiniteness assumption is dropped.Considerthe infimization of
(2.3)
overthe class ofinputsU(xo).
For agiven Xo aninputu*
is called optimal ifu*
U(xo)
andJ(xo,
u*)=
V+(xo).
Thefollowingwas proven in17].
THEOREM 4.1. Let
(A, B)
becontrollableandassumethatR >
O. Thenwehavethe following"(i)
V+(xo)
isfinite
for
allXoE
if
and onlyif
theARE
has a real symmetric solution(i.e., F
).
(ii)
lf
F
thenfor
allXo
,
V+(xo)=
xK+xo.
(iii)
If
F
thenfor
allXo thereexists anoptimalinputu*
if
andonlyif
A>
O. (iv)If
F
# and A>0 thenfor
eachXo
there is exactly one optimal inputu* and,
moreover, this inputu*
is given by thefeedback
control lawu*=
-R-(BrK++S)x.
As
already mentioned, an analogue of the latter theorem forthe free-endpoint case, up to now, has onlybeen available for the case that the quadraticformw(x, u)
is positive semidefinite, i.e.,forthe case that
w(x, u)>=O
for all(x,
u)["
m.
Inthe sequel,,letF+
:={K
6FIK
->_0}.
Itis well known[8],
12]
that ifto_>- 0 and if(A, B)
iscontrollable,
then theARE
hasasmallest.positive
semidefinite realsymmetricsolution.More
precisely, thereexists a (unique)K
such that(4.1)
/
F+,
(4.2)
K
F+
=:>
K
_<-K.
The solution
K
characterizedby(4.1)
and(4.2)
plays the central role in the solution of the free-endpoint regularLQ
problemwithpositivesemidefinite cost.In
the follow-ing,for agivenXo aninputu*
iscalledoptimalifu*
U(xo)
andJ(xo,
u*)
V+.(Xo).
THEOREM 4.2.
Assume
that(A, B)
iscontrollable,
thatR
> O,
and thatto(x, u)>-0
for
all(x, u)
’.
Then wehave the following:(i) For
allXo",
V.(Xo)
x(Xo.
(ii) For
each Xo,
thereisexactly
oneoptimalinputu*,
andmoreover,
this inputu*
isgiven by thefeedback
controllawu*
-R-(BT(
+
S)x.
Proof.
Thisfollows,
forexample, by combining 12,Thin.8]
andthe results from[1,
p.36]
(see
also[19]).
We
note that in this theorem the existence of optimal controls is no issue.In
contrast with the fixed-endpoint problem, the positive semidefiniteness assumption assures that in the free-endpoint problem for every initial condition there exists an optimal control.
In
tryingtogeneralize thelattertheoremtothecasethatw is anarbitrary indefinite quadratic formin(x,
u) (with
ofcourse,
asusual, R
>
0),
thefollowing aspects should be considered. First, due to the indefiniteness ofto,the optimal costVf(Xo)
nolonger needs to be finite.In
this paper we want to restrict ourselves to the case thatV(xo)
is finite forall Xo.
In
ordertoestablish aconditionassuring this,we state thefollowing well-known result.For
v0,
denotevl]
:=vrRv.
LEMMA
4.3.Let
K
F.
Thenfor
all uL2,1oc(+)
andfor
allT
>=
O,
wehaveJr(xo,
u)=
Ilu(t)+R-’(BK+S)x(t)[12n
dt+xroKxo-xT(T)Kx(T).
Here,
we have denotedx(t)
:=X(Xo,
u)(t).
Proof
For
aproof,
referto[2]
or17].
In
the sequel, let F_:=
{K
FIK
=<
0}.
From
the previous lemma the followingis immediate.LEMMA
4.4.Let (A, B)
becontrollable andR
>
O.If
F_ thenV(xo)
isfinite
for
allXo".
Proof.
I’_ implies thatK-=
<0. Applying the previous lemma toK-
yieldsJT(Xo, tl)
xoK-xo
forall u andT=
>0.Remark 4.5.
In
[17]
it issuggested
that the converse ofthe above lemma alsoholds,
i.e.,that finitenessofV.(Xo)
forall Xoimplies thatF_.
We
wereable neither to establish a proof nor to construct a counterexample to this assertion.We
were,however,
able to relate the condition F_ to an equivalent one in terms of the quantitiesJT(Xo,
U)
in a slightly different way.Indeed,
if(A, B)
is controllable andR >
0then the following equivalencecanbe proven:(4.3)
F_#:>
inf{liminfJT(Xo,
u)lu L2,toc(+)}
is finite for all Xo i".Tco
Note
thatif we could prove the above equivalence withL2,oc(+)
replaced byU(xo)
we would be done.Indeed,
foruU(xo)
we haveliminfT-ooJT(Xo,
U)=
IimT-JT(Xo,
U)--J(xo,
u),
so the infimum in(4.3)
would then be equal toV(xo).
conditions
for
the finiteness ofV.
remains a difficult open problem(see
also[18],
[11],
and[13]).
A
final point we want to make here is that for the free-endpoint problem with indefinitecost, even if theoptimal costis finite for all initial conditions,it is nottrue in general that optimal controls exist for all initial conditions.We
will illustrate this in the example below.It
should therefore be clear that part of our problem is to formulatenecessaryand sufficient conditionsfor theexistence ofthese optimal controls(as
was also done inTheorem 4.1(iii)).
Example4.6. Considerthecontrollablesystem -x
+
u,
x(0)
Xo with indefinite cost functionalJ(xo,
u)
-x(
t)
2+
u( t)
dt,
that is,take
A
1,B
1,Q
=-1,S
0,andR
1.The correspondingARE
isgivenby-2K-K2-1
=0. Consequently, K-=K/=-I.We
claim thatV.]:-(Xo)=-Xo.
We will showthis "fromfirst principles."Let
uEL.loc(R/).
For
every T-_>0we havex
+
u2dt(x
u)
dt+
2x(-x
+
u)
dt(x-u)
dr+2 xdt=(x-u
dt+x(T)-x.
Consequently,
J(xo,
u)-x
for all uU(xo).
On
the otherhand,
for e>0 define u(1
e)x.
Then ex andIt
followsthatVj(xo)=inf{J(xo,
u)lu
eU(xo)}
-xg.
Thus,
we see thatthe optimal cost is finite(as
could also be deduced from theNct
that K-=-1NO).
We
claim,however,
that noopimal control exiscsIndeed,
assumeu*
is optimal.Let
x*
be the correspondingtrNectory.
We
have-xg
J(xo, u*)
-xg
+
lira(x*
u*)
dt+
x*(
T)
T
From
this itfollows thatIo
(x*-u*)
dt 0andthat,
consequently,u*
x* However,
usingthisfeedbackcontrollaw yields
J(xo, u*)
0.Ifxo
0 thisyieldsacontradiction. 5. The free-endpoint regular LQ-problem with indefinite cost.In
this section we will resolve the free-endpoint version ofthe regularLQ
problemwith indefinite cost functional.In
the sequel, an important rolewill be played by the subspace(5.1)
N :=
(ker
K-IA->
X+(A-).
By
definition ofA-
it is immediatelyclearthat,
infact,
(5.2)
N
(ker
K-[A-BR-S)f’IX+(A-BR-S).
Obviously,
N
is equal to the undetectable subspace of(K-,
A-)
with respect to the stabilitysetCg
C-t.J
C.
We
alsonotethatN
is an A--invariantsubspace ofX/(A-).
By
Theorem 3.1,N
corresponds to a real symmetric solution),(N)
oftheARE. Let
PN
bethe projectorontoN
along A-N
.
Thenthis solution3,(N)
is given byItwillturnout that
K.t
+.,
the solution of theARE
supported bythe subspaceN,
isthe bottleneckinthe problemwe wantto resolve. We will show thatthe optimalcost for thefree-endpointproblem is obtainedfrom K.r+ and that theoptimal controls ifthey exist,are givenbythe feedback control law u=-R-(BTK+
.r+
S)x.
Before stating the exactresult we firstgive anintuitiveargumentas toexactly whythe subspace N given by(5..1)
is the "right" supporting subspace. The argument is as follows. First recall that ifw>=
0,then the optimalcostforthe free-endpointproblem is obtainedfrom the smallest positive semidefinite solution of theARE (see
Theorem4.2).
Now,
it canbe shownthat,
againifto>=
0, K3’(V)
is positive semidefinite if andonlyif VckerK-(see
Theorem6.2).
Consequently, ifto_-> 0 then the optimal cost is obtainedfrom the smallest solutionK
y(V)
of theARE
such that VckerK-Now,
our choice to consider exactly the subspace N given by(5.1)
is based on the guess that the latter statement is also valid ifto is indefinite.Note
that K.r+ is indeed the smallest solutionof
ARE
for which its supporting subspaceis contained inkerK-"
ifK y(V) is such that Vc kerK- then,
since V is A--invariant, .we must have Vc(ker
K-]A-)
(the
latter being the largest A--invariant subspace in ker
K-).
Also, VcX+(A-).
Thus,Vc N. Then it follows fromTheorem 3.2 that K+ <K. The following theorem is the main result of this paper.
THEOREM 5.1. Let
(A,
B)
becontrollableandassumethatR
>
O. Then wehave the following"(i)
V.;.(Xo)
isfinite
for
allXoR
if
theARE
has a negativesemidefinite
real symmetric solution(i.e.,
F_#).
(ii)
If
F_ (g thenfor
all Xo",
V.(Xo)
+ XoTK.t.
+Xo. (iii)If
F_ thenfor
all Xo"
there exists an optimal inputu*
if
and onlyif
kerA
ckerK-(iv)
If
F_ (g andif
kerA
kerK-,
thenfor
eachXoN"
there is exactly one optimal inputu* and,
moreover, this input is given by thefeedback
control lawu*=
_R-’(BTK
..
++
S)x.
In theremainderofthis section we will establishaproof ofthistheorem. Inorder tostreamline thisproof,we will statesomeof themostimportant ingredientsasseparate lemmas.
In
the firsttwolemmas,
we will formulate somegeneral structuralproperties of linearsystems.LEMMA5.2. Consider the system 2
Ax
+
,,
y Cx.Assume
thatC, A)
is observ-able. Let,
L2(+),
yLo(R+).
Thenfor
every initial condition Xo wehave xL(N+).
Proof
Since(C, A)
isobservable there existsamatrixLsuchthato’(A
+
LC)
C-Obviously, x satisfies the differential equation2
(A
+
LC)x
Ly
+
,,
x(O)
Xo..Usingthe variations of constants
formula,
togetherwith somestraightforward estimates, it is then easily verified that xUsingthe previouslemma we arriveatthefollowingresultthat will be one of the main instruments inthe proofof Theorem 5.1.
LEMMA
5.3. Consider the systemAx
+
,,
y Cx. Let Cgbea symmetric subsetof
C.Assume
that(C, A)
is detectable with respect toCg.
Let the state space be decomposedintoNn
X
X2,whereX1
isA-invariant.In
thisdecomposition,let x().
Assume
thatcr(A[X)
Cg
ando(AIN"/X)c
Cb
Thenfor
every initial condition Xo we have"if
,
L2(N
+)
andyL(N
+)
then x2Lo(N+).
Proof
We
claim that, infact,
X1
Xg(A).
Indeed,
the fact thatX
Xg(A)
is immediate.Denote
ro:=r(A]Xg(A)/X).
ThenCroCr(AlXg(A))cC.
Also, roCBy
thefact that(C, A)
isdetectable withrespectto Cgwemaytherefore conclude that(ker
C[A)c Xt.
Decompose
X
XX2,
withXll
:=(ker
CIA)
andX12
arbitrarily. Accordingly, letxl=(’,, We
XI21, then haveRn=XX2X2
withx=(x,x2, x[)
rIn
this decomposition, lettAl
A,2
A3
/
A
A22
A23
C--(0,
C2,
C3)
P P120
A33
/
P2Obviously, the system
C),
(A22
A23
0
A33])
isobservable.Moreover,
(2,2"]
2
/=(A2
0A33,]
\x2] P2 y(C2,
C3)(
P12.
2/It
thus follows fromLemma
5.2 that(X12’C L(R
+)
which of course implies that x2eLoo(N+).
E1Another important instrument in the
proof
that we will establish is the following result.LEMMA
5.4. Consider the system 2Ax
+
Bu,
x(O)
Xo.Assume
that(A,
B)
is controllableando’(A)
cC-U
C.
Thenfor
alle>
0 thereexistsacontrol uL2(
+)
such thatIo
Ilu(t)ll
dt<e andX(Xo,
u)(t)-O(t).
Proof
For
the given system considerthe fixed-endpoint regularLQ
problem inf{
f
llu(
t)l[2 dt]u
L(
+)
andx(xo, u)(
t)
O,
c}.
It
iswell known(see
also Theorem4.1)
thatthe above infimum is equaltoxK+xo,
whereK
/isthe maximal solution to the
ARE:
ArK
+
KA--KBBrK.
We
claim thatK
/=0.Assume K
/0. Since
K
0 is asolution to theARE,
we must have 0=<
K /.
So,
K+-
>0 andK+
0. Consequently, there exists an orthogonal matrix S such thatwith
K
>
O.Denote /
:=SK+S
,
:=SAS
r,
:0
:=SB.
Then we haver/ +//
KBB
rg.
Decompose
A2
A22]
B2
It
is easily seen thatA(.K+KA=KBIBK.
Also,
KIA2--0.
Since K>0, this impliesA2=0.
Define P:=K
-.
Then P>0 and satisfies theLyapunov
equationPA
+
AP
BB(.
Since(A, B)
iscontrollable,
thisimplieso-(A)
C+(see,
e.g.,[21, Lemma 12.2].
This,however,
contradicts the factthattr(Al)
tr(A)
tr(A)
C-LJ
C
.
We
conclude that the above infimum is zero. ElWe
have now collected the most important ingredients we need in the proofof our main theorem.In
order to give thisproof,
we shall make a suitable direct sumdecomposition of
the
state space. Let K.l+ be the solution of theARE (3 1)
defined by(5.3)
Denote
A;
:=A
BR-
(B
’K
.f++
S).
By
Theorem 3.1 we haveX
+(A,.)
+N,
X(A].)
kerA,
X-(A-;)
X-(A+)
fqA-’N
-.
Define
X,
:=X+(A)
X2
:=X(A-),
andXs
:=X-(A
+.f)
ThenX1
X2
Xs.
SinceXI
is A--invariant and sinceX2
is also A--invariant(ker
z
X(A:)
for allK
F)
we have
(5.4)
A-
A22
A23
0
A33
forgivenmatrices
Ai2.
We
also haveK.rx
+K-x
for allxN,
andhenceA;[X, A-IX,.
Also, since kerAcA-N
+/-and therefore
kerAckerPN,
for all xkerA we haveKx
K+x
K-x.Hence
A;[X:
A-[X2.
Consequently,0
(5.5)
A;
A22
for a given matrix
A3.
Note that g(A)m C+,
(A::)
C and(A3)
C-. SinceX
kerK-
andK-
is symmetric,0
(5.6)
g-=g
K5
Furthermore,
we claimthat A has the formA- 0
0
A33
Indeed,
byTheorem 3.1 we haveX2@ X3
zX-X
+/-
andthereforewemusthaveA13--0o
The other zero blocks are caused
.
by the fact thatX2
ker and by thesymmetry of Combining the representations for K- and,
we find+ 0 0 11
K
+ 0K
+ + 22K23
0 K+23K33
+ +for given matrices K+
(note
that, infact,
K:3
K3
andK22
K2).
Combining all this, we find that0
(5.7
*
j-K
;3We
now proceed with the following lemma, which states thatK
gives alower bound forthe optimal cost ofthe free-endpoint regularLQ
problem.LEMMA
5.5.Assume
that(A, B)
iscontrollable, R >
0, and F_#(.For
all Xon
and
for
alluU(xo)
wehave/
Io
(5.8)
J(xo,
u)>=xoK.rxo
+
]]u(t)+R-’(BK
+f
+
S)x(t)[]
dt.Herewehave denoted
x(
t)
:=X(Xo,
u)(
t).
Proof
Since F_ we have K-0. Let uU(xo).
It follows from Lemma 4.4 thatJ(xo, u)
is either finite or equal to+.
Indeed, J(xo, u)=-
would imply V.t.(Xo)
+=-,
which would contradictF_#.
Ofcourse, ifJ(xo, u)=
+
then(5.8)
holds trivially.Now
assume thatJ(xo, u)
isfinite.
By
the fact that K-0 it follows fromLemma
4.3 that for allT
0]]u(t)+R ’(BK
+S)x(t)[[
dtJv(xo, u)-xoK
Xo.0
Denote
(t)
:=u(t)+R-’(BK-+S)x(t).
It
thenfollowsthato
[(t)]]
2R
dt<
+,
and hence thatL2(+).
Again using Lemma4.3 and the fact that-K-
0,we findthat thisimplieslimwX(T)K-x(T)
exists(and
is finite). ThusK-x
must be bounded on+.
Denote
y(t):=
K-x(t).
SinceAx
+
Bu,
we have thatx,
u,andy arerelated by the equationsA-x
+
Bu,
yK-x.
Now let
"
be composed into"=
XX2X3
as introduced above. Write K-=(O,K,Kf),
B=(B(,Bf,
Bf)
andx=(x(,x[,x[)
SinceX,=
N
is theundetect-able subspace (with respect to
C-U
C)
of(K-,
A-),
it is easily verified thatthe pair K,
g),
0
A3/
is detectable
(with
respect toC-U
C).
Since(A-)
cC+U
C andsince
X
X(A-),
it can be verified that
0
A
c
Hence,
(A)c
C and(A)c
C+.
Also,we haveB
Since u
L(R
+)
and yL(R+),
byLemma
5.3 (applied withC
=C-U
C)
we have thatx
eL(R+).
Again by applying Lemma 4.3, this time with
K
Kj., we find that for allT
0(5.9)
JT(Xo,
u)=
Ilu(t)
+ R-,(BTK
..
+ T + T+S)x(t)l[ dt+xoKxo-x (T)Kf
x(T).
Denote
w(t)
:=u(t)+
R-(BTK
+f
+
S)x(t)
Combining(5.6),
(5.7)
and(5.9)
weobtainthat for all
T
0T
T +
(5.a0) JT(Xo, U)=
Ilw(t)ll dt+xoK..Xo-
xf(T)x(T)-xT(T)K-x(T).
Recall thatlimT
JT(Xo,
U)
was assumedto be finite.Thus, JT(Xo, U)
is a bounded function ofT.
Sincealsox3(T)
andxT(T)K-x(T)
areboundedfunctions ofT,
(5.10)
We
againconsider(5.10).
SincenowJr(xo, u),
Io
IIw(t)[l
dt andxT(T)K-x(T)
converge for
T
-
oe,
it follows that limr_,xr(T)33x3(T)
exists. Since33
>
0 this impliesthat1123(T)1
converges asr
.
Now,
since2Ax
+
Bu,
the variablesx and w arerelatedvia2Afx
+
Bw,
and hence(see
5.5)
2A;3x
+ B3w.
Since weL2(N
+)
and
(A;3)
cC-wehave thatx3L2(+).
Afortiori, since[[x3(/)ll
convergesas,
thisyields lim,x3(t)
0. Using this, and thefact that-K-
0,itthen follows from(5.10)
that(5.8)
holds.S
Our next lemma states
that,
by choosing the control properly, the difference between Kf+ and the value ofthe cost functional canbe made arbitrarily small.LEMMA 5.6.
Assume
that(A,
B)
iscontrollable, R >
O,
andF
.
en
for
allT +
Xo N and
for
alle>0 thereexists an inputuU(xo)
such thatJ(xo,
u)
N XoK
fxo
+
e.Proo
Again, letN"
bedecomposed asabove.It
follows from(5.7)
and(5.9)
that for all uL,o(+)
andfor allT
0r +
-(x(T),x(T))
K:?
K33
]
xg(T(s.)
J(xo,
)=
ll(t)ll
dt+xoK..Xo
+Here,
w:=u+ R-(BrK
+f
+
S)x.
Since 2Ax
+
Bu,
the variables x and w are relatedby2=
Ax
+
Bw,
and hence(see (5.5))
Note
that(Ae)
cC,
(A;3
cC- andthatthissystemiscontrollable.Now
let e>
0. It follows fromLemma
5.4 that there exists a controlwL(N
+)
such thatIo
[[w(t)lldt<
and such thatx(T)O
andx3(T)0
as T. Defineu:=
_R-I(BrK
+f
+
S)x
+
w. Then we haveT + T +
J(xo, u)=
limJr(xo,
u)=
Ilw(t)ll
dt+xoK .Xo +xog. .Xo.
T
We
will now prove our maintheorem.Proof
of
Theorem 5.1.(i)
Thisproofwas already statedseparatelyinLemma
4.4.(ii) Lemma
5.5yieldsJ(xo, u)
>-XorKr
+Xo for all uU(xo).
TogetherwithLemma
5.6 this implies
Vt
+-
(Xo)
x
rK
f+Xo for all Xo.(iii) Assume
F_#.
(3)
Assume
that for all Xothereexists acontrolu*
U(Xo)
such thatJ(xo,
u*)
Vf.(Xo)
T +=XoKrXo.
Let
XoR
be arbitrary and letu*
be thecorresponding optimal control.
Denote
x*
:=X(Xo,
u*).
By Lemma
5.5T +
=J(xo,
u*=
T +*(
TK+
xoKfxo
)>xoKfxo+
Ilu
t)+R- (B
f+S)x*(t)llat.
It
followsthatu*
must begiven by the feedbackcontrol lawu*
R-I(B
rKf/
+
S)x*.
This implies thatx*
satisfies the equation2*=
Af.x*.
In
terms ofthe decomposition introduced above, this of course yields22* =A22x*2
and23*
A33x
*
(see
5.5).
Sinceo’(A3
cC- we musthavex3*(t)
-
0(t
-->oo).
By
(5.10)
J-r(Xo, u*)
T +Xo
Kfxo
x*3r(T)A3x*3(T)-x*r(T)K-x*(T).
T +
By
the fact thatJr(xo, u*)-
XoKfxo
we obtain thatx*
Since K-issemidefinite, afortiori this implies
K-x*(T)-O (T- oo).
Using(5.6)
this yieldsSince
x*3(T)O (T-->)
the latter impliesK2x*2(T)-->O
(T-->o)
or, equivalently,K2
exp(A22T)x2(O)->O (T-->). Now, x2(0)
was completely arbitrary andtherefore we find thatK,2 e A22T-->0
T
-->Consequently,
K2(Is-A22)
-1has all its poles in
C-. On
the otherhand, however,
since
or(A22)
cC,
ithas all itspolesinC.
Thus,
K2(Is
Az)
-10,andhence
Kz
0. SinceK-
is semidefinite thisimpliesK-3
0. Itfollowsthat kerAX2
ckerK-.
()
Conversely, assume kerA
kerK-.
ThenK
=0 andK3
=0. Define u-R-(BTKf+S)x.
We
claim that this feedback law yields an optimal u.Indeed,
by
(5.11)
J(Xo, u)
xoKyxo
+x(T)
+K
33X3(Z).
Moreover,
23
At33x3
Sinceo’(A3
cC- we havex3(T)
-
0(T--> c).
ThusJ(xo,
u)
T + +
Xo
K.t
XoV.f
(Xo),
so u is optimal.(iv)
Thefact thatu*---R-(BrK.
+
S)x*
is unique was already proven in(iii)
(3).
This concludesthe proof ofourtheorem. E]Remark5.7.
At
thispointwewould like to mentionthat,
in addition totheoption wehavechosen in 2,there is still anotherverynaturalandappealingway to formulate the regularLQ
problem. Instead ofrestrictingthe class of controls toU(xo)
in order to guaranteethatthe indefinite integrals in(2.2)
arewell-defined, itis also possibleto chooseL2.1oc(R/)
for the class of admissible controls and to consider the following cost functional:,(Xo,
u):=
lim supJ-(Xo, u).
Tx
Obviously, on the subclass
U(xo)C
Lz,oc(R+)
the functionalsJ(Xo,’)
andJ(xo,’)
coincide. Corresponding to this choice ofcost functional, we can now consider the following version of
the
free-endpoint regularLQ
problem:’f.
(Xo):=
inf{J(Xo,
u)lu L2,1oc([+)}.
As
it turnsout,wecandevelop around this versionof theproblematheory completely parallelto the one we developed in this section. Infact,
Theorem 5.1 remains valid if in its statement we replaceV/.t
byQ+.t.
In particular, both problems yield the same optimal controlsu*.
Consequently, ifu*
is optimal for the problem with functionalJ(Xo,"
),
then in factu*
U(xo)
and9i.(Xo)= J(Xo,
u*)=
limT-_JT-(Xo,
u*).
Similar remarks hold for the fixed-endpoint problem.6. Comparison and special eases.
In
this section we will discuss some questions that ariseif we compare the optimal costs and optimal closed loop systems resulting fromthe free-endpointandfixed-endpoint problem, respectively.In
particular,we will establish conditions under which the respective optimal costs are the same.Also,
conditions will befound underwhichthe free-endpoint optimalclosed loop system is asymptotically stable. Finally,we willshow howourgeneralresults can be specialized to reobtain the most important results on thefree-endpointregular
LQ
problemwith positivesemidefinite
cost functional. First, we havethe followingtheorem.THEOREM
6.1.Assume
that(A,
B)
iscontrollable, R >
O,
andF
Q. Thenwehave thefollowing:(i)
K.)+:=
K
/if
and onlyif
the pair(K-,
A-BR-*S)
is detectable with respect to the stabilityset C-[_JC.
(ii)
o-(A.-)c
C-if
and onlyif
the pair(K-,
A-BR-S)
is detectable with respect to C- andA >
O.Proof
(i) By (5.2), N
is equaltotheundetectable subspace of(K-,
A-BR-1S)
withrespectto
C-U
C.
Since K+ is supportedby thezero subspace, by Theorem 3.1+ g+
we have Ks ifand onlyif N 0.
(ii)
(=)
Detectability with respect to C- implies detectability with respect to+ + + +
C-
U
C.
Hence K
sK
andA
sA
By
17,Thm.5]
A>
0 ifandonlyifo-(A
+)
cC(=:>)
Conversely,assumecr(A-)=
C-By [17,
Thin.5]
there is exactlyoneKeF,
+ +
A+=
A
+.
namely
K
K+,
suchthato-(AK)
=
C-U
C.
Hence K
sK
s Consequently,ZX
>
0.Also,
from(i)
we obtainthat the pair(K-, A
BR
-
S)
isdetectablewithrespect toC-UC
.
SinceA>0, o-(A-)
cC+.
Hence
X(A-)
0 so(K-,A-BR-IS)
isin fact detectablewith respect toC-.We
will now discuss how ourresults canbe specializedtorederive someimportant "classical"resultsonthe specialcase thatthe quadratic formtoispositivesemidefinite.We
have the following characterization of the positive semidefinite solutions oftheARE.
THEOREM 6.2.
Assume
that(A, B)
iscontrollable, R >
O,
F_,
andF+
.
LetK
F
be supportedby
V. ThenK
F+
if
andonlyif
VckerK-.Proof
By
Theorem 3.1 we haveV@
A-
V-
N".
() Assume
that Vcker K-.ThenA-IV{x
Nnly
rK+x
0,forallyV}
andK
K+(I
Pv).
Let
x[n, x=x+x2
withx
V andx2A-W
-.
It
is easily seen thatxrKx=xfK+x2.
SinceF+
we haveK+->0.
It
follows that K=>0.()
Conversely,ifK
_>-0 then for all x Vwe have0
<=
xrKx
xr
(K-Pv
+ K+(I
Pv))X
xrK-x.
Since F_ wehave K-_-<0.
It
followsthatxrK-x
0,and hencethatxOur next result states
that,
under the assumption that F_,
if theARE
has positivesemidefinitesolutions atall,then it hasasmallestpositive semidefinitesolution and this solution is equal to the one supported by N.THEOREM 6.3.
Assume
that(A, B)
iscontrollable, R > O,
and F_.
Then the following hold"if
F+
# then(i)
Ks+F+
and (ii)KF+
impliesKs+K.
Proof
Since NckerK-
itfollows from Theorem6.2
thatK.t
+.
F+.
Now
assumeK
F+
and K is supported bythe A--invariant subspace VX+(A-).
Since K F/ we haveV
ckerK-. Hence
Vc(ker
K-[A-)
(the
latter is the largest A--invariant subspace in kerK-;
see[21]).
It follows that VcN.
But then, by Theorem 3.2,+<K.
Ks--From the above we deduce the following remarkable fact. Consider the free-endpointregular
LQ
problemwithindefinite
costfunctional.Let(A, B)
be controllable.We
already saw thatthe optimal cost is finite if we have F_.
Assume
this to be the case. Then Theorem 6.3 states thatif
theARE
hasatleastonepositivesemidefinite
solution, then the optimalcost isgiven by the
smallest
of
these solutions! The case that thecost
functional is positive semidefinite, i.e.,to(x, u)>-_ O,
for all(x, u),
is in fact a special case of this general principle.Indeed,
if(A,
B)
is controllable and if to_->0 thenF+
(see
[5]). Moreover,
applying the lattertothecontrollablesystem(-A, -B)
and the same form to_->0, we
can
also see that F_.
Thus we have reobtained Theorem 4.2(i).Our
nextresult shows that the fact that for the case to-> 0 optimal controls exist for all initial conditions is also a special case ofamore general principle.PROPOSITION6.4.
Assume (A, B)
iscontrollable, R >
O,
F_,
andF+
.
Then ker A kerK-.
Proof
F_ isequivalenttoK-
0 andF+
is equivalent toK
+=>
0.Assume
x ker
A.
Then 0 __-<xrK
+x
xTK-x
_--<0.Thus
xrK-x
0, and henceK-x
0.By
combiningthis with the above remarks andby applying Theorem 5.1(iii) and(iv)
we reobtain Theorem4.2(ii).
To
concludethis section,we willbrieflydiscuss what statements can be obtained from Theorem 6.1 forthe case that our cost functional is positive semidefinite.In
the restofthis section, assume thatto(x, u)>=_
0 for all(x, u). We
claim that in this case(6.1)
N(ker Q
STR-’S)IA
BR-’S)
CI
X+(A
BR-’S).
First we claim that kerK- is
(A-BR-S)-invariant.
Indeed,
if w>=0 thenQ-STR-S>=
O. Also it is straightforward to verifythat(6.2)
(A-BR-’S)TK-+K-(A-BR-’S)+Q-STR-’S-K-BR-’BTK=O.
Let Xo kerK-.Thenfrom
(6.2),
xTo(Q
STR
-S)xo
0,andhence(Q
STR
-
S)xo
O.Thus,
again from(6.2),
K-(A-BR-S)xo=O
so(A-BR-S)xoker
K-. It followsthat
(ker
K-IA-BR-S)
kerK-.
Now,
by using the interpretation of K- as the optimal cost fora fixed-endpointLQ
problem in "reversed time"(see [21,
Thm.7])
it canbe provedthatker
K-
(ker Q
STR -1S)IA BR-’
S)
(6.3)
Ci
(X+(A
BR-’
S)@
X(A
BR
-lS)).
Thus
(6.1)
follows immediately from(5.2) We
have now shown that ifw>0,thenK+ isin factsupported bytheundetectablesubspaceofthepair(Q-STR-S, A-BR-S)
with respect toC-LJ
C.
(See
also[3,
Thm.1].)
By
applying Theorem6.1(i)
we may+
then concludethat
KT
K
if andonlyif(Q
STR
-S,
A
BR
-S)
isdetectablewith respect toC-
CO(see
also[12,
Cor.,
p.356]).
Finally, we will re-establish the well-known fact that
o-(A.-)c
C- if and only if(Q---SrR-S, A-BR-1S)
is detectable with respect to C-(see
[6],
[20],
and[12]).
Assume
thatto=>0.We
claimthat if(K-,
A-BR-1S)
is detectablewith respectto C-then A>0.Indeed,
if(K-,
A-BR-S)
is detectable with respecttoC- then(K-, A-)
is detectable with respect to C-. The latter isequivalentto
(6.4)
(ker
K-IA-)(-I
(X+(A-)X(A-))=0.
By
Theorem 3.1,X(A
-)
kerA. Also,
since co_>-0, kerZX
ckerK-.Hence,
by(6.4),
kerz+((kerK-[A-)OX+(A-))=O,
whence ker=0.It
follows that A>0.We
maynow concludefromTheorem
6.1(ii)
thatcr(A)
C- ifand onlyifthe pair(K-,
A-BR-S)
is detectable with respect to C-From
the fact that kerK-
is(A-BR-S)-invariant
and from(6.3),
the latter condition is,however,
equivalent to the statementthat the pair(Q-SrR-S, A-BR-S)
is detectable with respect to C-.7. Concluding remarks.
In
thispaperwe have studiedjustone of themany open basic questions that still exist in the context oflinear quadratic optimal control. To namebut afew ofthese open problems,we mention,forexample, the questionabout the relationship between thefinite-horizon
free-endpoint problem and the infinite-horizonfree-endpoint problem.It
is wellknownthat if the cost functional ispositive semidefinite, then the finite-horizon optimal cost converges to the infinite-horizon optimal cost[1], [2], [9].
It would be interesting to investigate whether this is also true for the indefinite case. Another open problem is the singularLQ
problem with indefinite cost functional, that is, theproblem
studied here without the assumption thatR
is positivedefinite. Recently[19]
thisproblemwas treatedfor thecase thatthecost-functional is positive semidefinite.
However,
for both the free-endpoint case as well as the fixed-endpointcase, the indefinite versionof this problemstill remains to be solved.Acknowledgments.
I
thank Dr.Jacob van derWoude and ProfessorMaloHautus
forsomevery usefuldiscussions whiletheresearchleadingtothispaperwas carried out.
REFERENCES
[1] B. D. O. AqDERSON ANDJ. B. MOORE, LinearOptimalControl, Prentice-Hail, Englewood Cliffs, NJ,1971.
[2] R. W. BROCKETT,FiniteDimensionalLinearSystems,John Wiley,New York, 1970.
[3] F. M. CALLIERANDJ. L. WILLEMS,Criterionforthe convergenceofthe solutionoftheRiccatidifferential equation, IEEETrans.Automat.Control, 26(1981), pp. 1232-1242.
[4] W.A.COPPEL,Matrixquadratic equation, Bull. Austral. Math.Soc., 10(1974), pp. 377-401.
[5] R. E.KALMAN, Contributionstothetheoryofoptimalcontrol, Bol.Soc.Mat.Mexicana(2),5(1960), pp. 102-199.
[6]
,
When isalinearcontrol systemoptimal? Trans.ASME J.BasicEngrg., 83(1964), pp. 51-60.[7] V.KUCERA,Acontributiontomatrixquadratic equations,IEEETrans.Automat.Control,AC-17(1972), pp. 344-347.
[8]
,
On non-negative definite solutions to matrix quadratic equations, Automatica, 8 (!972), pp. 413-423.[9] H. KWAKERNAAKAND R. SIVAN, LinearOptimalControlSystems, Wi!ey-lnterscience, New York, 1972.
[10] K.MARTENSSON,On thematrixRiccatiequation, Inform.Sci., 3 (1971),pp. 17-49.
[11] B. P. MOLINARI, Conditionsfor nonpositive solutions ofthe linear matrix inequality, IEEE Trans. Automat.Control, 20(1975),pp. 804-806.
12]
,
Thetime-invariantlinear-quadratic optimalcontrolproblem,Automatica, 13(1977),pp. 347-357.[13] P.J.MOYLAN,Onafrequencydomain conditioninlinear optimal controltheory,IEEETrans.Automat.
Control,29(1975), p. 806.
[14] V. M. PoPov, Hyperstability and optimalityofautomaticsystems with severalcontrolfunctions, Rev. RoumaineSci.Tech.Set.Electrotech.Energet.,9(1964),pp. 629-690.
[15] J. M. SCHUMACHER, Dynamic Feedbackin Finite andInfinite Dimensional LinearSystems, Math. CentreTracts, 143,Amsterdam,theNetherlands, 1981.
[16] M.SHAYMAN,Geometryofthe algebraicRiccatiequationmpart 1,SIAM J.Control Optim., 21 (1983), pp. 375-393.
[17] J.C.WILLEMS, Leastsquares stationaryoptimal control andthealgebraicRiccati equation,IEEETrans.
Automat.Control, 16 (1971),pp. 621-634.
[18]
,
Ontheexistenceofanon-positive solutiontotheRiccatiequation,IEEE Trans. Automat.Control, 19(1974), pp. 592-593.[19] J.C.WILLEMS,A. KITAP(I,ANDL. M.SILVERMAN,Singular optimalcontrol,ageometricapproach, SIAM J. Control Optim., 24(1986),pp. 323-337.
[20] W.M. WONHAM, Ona matrixRiccati equationofstochasticcontrol, SIAM J. Control Optim.,6(1968), pp. 681-698.