Preprints, 1st IFAC Conference on Modelling, Identification and

Preprints, 1st IFAC Conference on Modelling, Identification and

Preprints,
Control of 1st IFAC Conference
Nonlinear Systems onon Modelling, Identification and
Preprints,
Control of 1st IFAC Conference
Nonlinear Systems Modelling, Identification and
June 24-26,
Control
Control of 2015. Saint
Nonlinear
of Nonlinear Petersburg,
Systems
Systems
Available online at www.sciencedirect.com
Russia
June 24-26, 2015. Saint Petersburg, Russia
June
June 24-26,
24-26, 2015.
2015. Saint
Saint Petersburg,
Petersburg, Russia
Russia
ScienceDirect
IFAC-PapersOnLine 48-11 (2015) 238–241
Kalman—Yakubovich—Popov
Kalman—Yakubovich—Popov Lemma
Lemma And
And
Kalman—Yakubovich—Popov
Kalman—Yakubovich—Popov Lemma
Lemma And
And
Hilbert’s
Hilbert’s 17th
17th Problem
Problem
Hilbert’s 17th Problem
Hilbert’s 17th Problem
Sergei
Sergei V.
V. Gusev
Gusev
Sergei
Sergei V. V. Gusev
Gusev
St.Petersburg
St.Petersburg State
State University
University
St.Petersburg, St.Petersburg
St.Petersburg State
Universitetskaia
State University
nab. 7/9,
University 7/9, Russia
Russia
St.Petersburg,
St.Petersburg, Universitetskaia
Universitetskaia nab.
nab. 7/9, Russia
St.Petersburg, (e-mail: gusev@ieee.org).
Universitetskaia
(e-mail: gusev@ieee.org). nab. 7/9, Russia
(e-mail:
(e-mail: gusev@ieee.org).
gusev@ieee.org).
Abstract: Kalman-Yakubovich-Popov
Abstract: Kalman-Yakubovich-Popov
Kalman-Yakubovich-Popov (KYP) (KYP)
(KYP) lemmalemma
lemma is is the
the cornerstone
is the cornerstone of
of control
control theory.
theory. It
It
Abstract:
was used
Abstract: in thousands of
Kalman-Yakubovich-Popov papers in many (KYP) areas of
lemma automatic cornerstone
control.
is the cornerstone The of control
new
of new
control theory.
versions
theory.and and It
It
was
was used
used in
in thousands
thousands of
of papers
papers in
in many
many areas
areas of
of automatic
automatic control.
control. The
The new versions
versions and
generalizations
was used in
generalizations of
of KYP
thousands
KYP lemma
of
lemmapapers emerge
in
emerge in
many
in literature
areas
literature ofevery year.
automatic
every year. The
The original
control.
original The formulation
new
formulation versions of
of KYP
and
KYP
generalizations
lemma claims
generalizations of
of KYP
the lemma
equivalence
KYP lemma of emerge
three
emerge in literature
literature every
statements: 1) year.
year. The
fulfillment of original
so-called formulation
frequency-domain of
of KYP
lemma
lemma claims the
claims the equivalence
equivalence of
of three in
three statements:
statements:
every
1) fulfillment
1) fulfillment The oforiginal
of so-called
so-called
formulation
frequency-domain
frequency-domain
KYP
inequality,
lemma claims
inequality, 2)
2) solvability
the equivalence
solvability of
of the
the KYP
ofKYP linear
three linear matrix
statements:
matrix inequality,
1)inequality,
fulfillmentand and 3)
of so-called
3) solvability
solvability of the
frequency-domain
of the Lur’e
Lur’e
inequality,
equation.
inequality,The 2)
2) solvability
equivalence
solvability ofofthe
firstKYP
two linear
linear matrix
statements was inequality,
proved by and 3)
3) solvability
V.A.Yakubovichsolvabilityand of
of the
is Lur’e
further
equation.
equation. The
The equivalenceofof
equivalence ofthe
firstKYP
first two statements
two statements matrix wasinequality,
was proved by
proved by andV.A.Yakubovich
V.A.Yakubovich and
and
the Lur’e
is further
is further
called
called Yakubovich
equation. The equivalence
Yakubovich statement.
statement. of first two statements was proved by V.A.Yakubovich and is further
called
The
called Yakubovich
paper investigates
Yakubovich statement.
whether
statement. the KYP lemma holds when the field of real numbers is replaced
The
The paper
paper investigates
investigates whether the
whether the KYP
KYP lemma lemma holdsholds when
when thethe field
field of of real
real numbers
numbers is is replaced
replaced
by
by some
Thesome other
paperother ordered
investigates
orderedwhether field.
field. The
The necessary
the necessary
KYP lemma and
and sufficient
holds condition
when condition
sufficient the field ofisreal is found
found numbersfor
for Yakubovich
is replaced
Yakubovich
by some
statement other
by some other to ordered
hold in
ordered field.
ordered The
field. Thefields.necessary
It
necessaryis shown and sufficient
and that condition
Yakubovich
sufficient condition is
statement found
is foundcan for Yakubovich
for hold in
Yakubovich such
statement
statement to hold
to hold in ordered
in ordered fields.
fields. It is
It is shown
shown that
that Yakubovich
Yakubovich statement
statement can
can hold
hold in such
in such
fields
fields when
statement
when to Lur’e
hold
Lur’e equation
in ordered
equation (and
(and corresponding
fields. It is
corresponding shown Riccati
that
Riccati equation)
Yakubovich
equation) has
statement
has no
no solution.
can
solution. hold Based
Basedin on
such
on
fields
the
fields when
statement
when Lur’e
Lur’eof equation
Hilbert’s
equation (and
17th
(and corresponding
problem
corresponding it is Riccati
shown
Riccati thatequation)
if the
equation) has
matrices
has no
no solution.
in the
solution. Based
formulation
Based on
on
the
the statement
statement of Hilbert’s 17th problem it is shown that if the matrices in the formulation
of
of statement of
theYakubovich
Yakubovich of Hilbert’s
statement
Hilbert’sdepend
statement
17th
17th problem
depend rationally
problem
rationally
it
it onis
is shown
on parameters,
shown
parameters,
that
that if the
ifthen matrices
the there
then matrices
there exists
exists
in the
the formulation
in solution
solution of
formulation KYP
of KYP
KYP
of
of Yakubovich
inequality
Yakubovich which statement
is also
statement depend
rational
depend rationally
function
rationally of on parameters,
these
on parameters.
parameters, then
then The there
there exists
generalized
exists solution
formulation
solution of
of KYP of
inequality
inequality which
which is
is also
also rational
rational function
function of
of these
these parameters.
parameters. The
The generalized
generalized formulation
formulation of
of
Yakubovich
inequality
Yakubovich statement
which is
statement also and
and Hilbert’s
rational
Hilbert’sfunction17th
17th problem
of these
problem for abstract
parameters.
for abstract ordered
The
ordered fields
generalized
fields is
is presented.
formulation
presented. It
It is
of
is
Yakubovich
shown that
Yakubovich statement
generalized
statement and
and Hilbert’s
versions of
Hilbert’s 17th
Yakubovich
17th problem
problem for
statement
for abstract
and
abstract ordered
the
ordered fields
statement
fields is
of
is presented.
Hilbert’s
presented. It
17th
It is
is
shown
shown that generalized versions of Yakubovich statement and the statement of Hilbert’s 17th
shown that
problem
problem are
that
are
generalized
equivalent.
generalized
equivalent.
versions
versions of of Yakubovich
Yakubovich statement statement and and the the statement
statement of of Hilbert’s
Hilbert’s 17th 17th
problem
problem are
are equivalent.
equivalent.
© 2015, IFAC (International Federation of Automatic Control) Hosting by Elsevier Ltd. All rights reserved.
Keywords:
Keywords: KYP
KYP lemma, frequency
lemma, frequency
frequency domain domain inequality,
domain inequality,
inequality, LMI, LMI, Lur’e
LMI, Lur’e equation,
Lur’e equation,
equation, SOS, SOS, Hilbert’s
SOS, Hilbert’s
Hilbert’s
Keywords:
17th problem
Keywords: KYP lemma,
KYP lemma, frequency domain inequality, LMI, Lur’e equation, SOS, Hilbert’s
17th
17th problem
17th problem
problem
1. INTRODUCTION
1. INTRODUCTION
INTRODUCTION the Kalman—Yakubovich—Popov lemma and exact loss-
1. the
the Kalman—Yakubovich—Popov
Kalman—Yakubovich—Popov lemma
lemma and
and exact
exact loss-
loss-
1. INTRODUCTION lessness
lessness condition”
the Kalman—Yakubovich—Popov
condition” got
got best
best student
student paper
lemma
paper award
and
award on
exact
on 50th
loss-
50th
The Kalman—Yakubovich—Popov (KYP) lemma is a lessness
CDC.
lessness condition”
condition” got
got best
best student
student paper
paper award
award on
on 50th
50th
The Kalman—Yakubovich—Popov (KYP) lemma is a CDC.
The
The Kalman—Yakubovich—Popov
cornerstone of control
Kalman—Yakubovich—Popov theory. It (KYP)
emerged
(KYP) as lemma
aa tool isforaa CDC.
CDC.aim of this article is to explore the algebraic roots
cornerstone
cornerstone of control
of control theory.
theory. It It emerged
emerged aslemma
as a toolisfor
tool for The The aim of this article is to explore the algebraic roots
absolute
cornerstone stability
of analysis
control theory.of nonlinear
It emerged control
as a systems.
tool for The aim of
absolute stability
absolute
Since that
stability analysis
it found
analysis of
applications
of nonlinear
nonlinear control
in optimal control,
systems. of
control systems.
signal The
of the
theaimKYP
KYP of this
lemma.
this
lemma.
article
article While
While
is
is totothe
the
explore
KYP
explore
KYP
the
lemma
the
lemma
algebraic
was
algebraic
was
roots
always
roots
always
absolute
Since that stability
it found analysis
applications of nonlinear
in optimal control
control, systems.
signal of
of the
formulated
the KYP
KYP in lemma.
terms
lemma. While
of real
While the
and
the KYP
complex
KYP lemma
lemmanumbers,was
was always
we are
always
Since
Since that
that it
processing, found
found applications
itadaptive and robust
applications in
in optimal
control,
optimal control,
control
control, of signal
infinite formulated
signal considering
formulated in
in terms
terms of
of real
real and
and complex
complex numbers,
numbers, we
we are
are
processing, adaptive and robust control, control of infinite formulated in
considering the
the possibility
terms of real to
possibility to
and extend
complex
extend the
the lemma
numbers,
lemma to
toweother
are
other
processing,
dimensional adaptive
systems.
processing, adaptive and
The
and robust
KYP
robust control,
lemma
control,plays control
plays an
control of infinite
important
of infinite fields.considering the possibility to extend the lemma to other
dimensional systems. The KYP lemma an important fields. A
considering
A generalized
the
generalized formulation
possibility
formulation to of
extend
of the
the thelemma
lemma
lemma for
for ordered
to other
ordered
dimensional
role as systems. The KYP lemma plays an important
role
role as aaa mathematical
dimensional
as
mathematical
systems. The
mathematical bridge
KYPbetween
bridge
bridge
lemma plays
between
between
frequency
frequency
frequency
domain
an important
domain fields
domain
fields.
fields.
fields
A
is
is
generalized
A given.
generalized
given. The
The
formulation
paper
formulation
paper is
is based
based
of the
of on
theour
on
lemma
lemma
our
for
for ordered
previous
previous result
ordered
result
and
role
and state
as
statea space
space methods
mathematical
methods in
bridge
in control
between
control theory.
frequency
theory. domain fields
Gusev is given.
(2014), The
where paper
a is based
sufficient on
conditionour previous
was foundresult
for
and state space methods in control theory. fields
Gusev is given.
(2014), The
where paper
a is based
sufficient on
conditionour previous
was foundresult
for
and
The state
first space methods
formulation and in control
proof of theory.
the lemma was pub- Gusev
lemma (2014),
to
Gusev (2014), hold where
in abstract
where a sufficient
ordered
a sufficient condition
field.
condition Now was
we
was found
prove for
that
foundthatfor
The first formulation and proof of the lemma was pub- lemma
lemma to
to hold
hold in
in abstract
abstract ordered
ordered field.
field. Now
Now we
we prove
prove that
The
lished
The first
by
first formulation
Yakubovich
formulation and
(1962)
and proof
proof in of
of the
Russian
the lemma
lemma was
Mathematical
was pub-
pub- this
lemma
this condition
to hold
condition is
in
is in fact
abstract
in fact a
a necessary
ordered
necessary and
field.
and sufficient
Now we
sufficient one.
prove
one. This
that
This
lished byby Yakubovich
Yakubovich (1962) (1962) in in Russian
Russian Mathematical
Mathematical implies this condition aais in
lished
Doklady.
lished by Since
Doklady. than
Yakubovich
Since thousands
(1962) inof
than thousands
thousands papers
of Russian
papers have have used
Mathematical
have used thisthis this
implies
this statement
that
condition
that in fact
fact aa necessary
isstatement
statement of the
necessary
of the
and
and sufficient
lemma
lemma is
sufficient
is
one.
one. This
equivalent
equivalent to
This
to aa
Doklady.
result and Since
more than
then one hundred of papers
versions and used
generaliza- implies thatof a statement
generalized of
version the lemma
of Hilbert’s is equivalent
17th to
problem. a
Doklady.
result andSince
more than
then onethousands
one hundredofversionspapers and
versions used this statement of generalized version of Hilbert’s 17th problem.a
havegeneraliza- implies that a statement of the lemma is equivalent to
result
tions and more then hundred and generaliza- statement of generalized version of Hilbert’s 17th problem.
result was
tions and published.
was more then one
published. The original
Thehundred Yakubovich
original versions
Yakubovich paper
and generaliza-
paper was was
was statement of generalized version of Hilbert’s 17th problem.
tions
tions was
included
was published.
into book
published. The
”Control
The original
Theory:
original Yakubovich
Twenty-Five
Yakubovich paper
paperSeminalwas 2.
2. FORMULATION
FORMULATION OF THE PROBLEM.
included into
included
Papers
into book
book ”Control
(1932-1981)”
”Control Theory:
edited
Theory: Twenty-Five
by T.Basar.
Twenty-Five Seminal
It is
Seminal
interesting 2. FORMULATION OF
2. FORMULATION OF THE
OF THE PROBLEM.
THE PROBLEM.
PROBLEM.
included(1932-1981)”
Papers into book ”Controledited Theory:
by T.Basar.
T.Basar.Twenty-Five
It is Seminal
is interesting
interesting
Papers
to note
Papers (1932-1981)”
that besides
(1932-1981)” edited
Yakubovich by
edited by T.Basar. article It
this
It book includes
is interesting Let us begin from aa classical formulation of the Kalman—
to
to note that
note that besides Yakubovich
besides Yakubovich article this
article this book
book includes Let
includes Let us
us begin
begin from
from aa classical
classical formulation
formulation of
of the
the Kalman—
Kalman—
three
to more
note that papers
besides related
Yakubovich to the KYP
article lemma.
this book includes Yakubovich—Popov
Let us begin
Yakubovich—Popov from lemma
classical
lemma Gelig et
formulation
Gelig et al.
al. (1978).
of the
(1978). Kalman—
three
three more
more papers
papers related
related to
to the
the KYP
KYP lemma.
lemma. Yakubovich—Popov lemma Gelig et al. (1978).
three
The more
interest papers
to the related
KYP to
lemma the KYP
has not lemma.
vanished in recent Yakubovich—Popov
Theorem 1. Let A ∈ R
R
n×n
lemma , B ∈ R
Gelig n×1et al.
and (1978).
G ∈ SM n+1 (R). If
n×n, B ∈ Rn×1 and G ∈ SMn+1 (R). If
n×n n×1
The interest
interest to to the
the KYP
KYP lemma lemma has has notnot vanished
vanished in recent Theorem
in recent Theorem 1.
1. Let
Let A
A ∈
∈ R n×n, Bthen ∈ R and G ∈ SM
The a pair
Theorem (A, B)
1. is
Let controllable,
A ∈ R , B ∈ R the
n×1 following
and G ∈ SM n+1 (R).
statements If
years. The
The interest paper
to theby Iwasaki
KYP lemma and Hara
has not(2005)”Generalized
vanished in recent aa pair (A, B) is controllable, then the following n+1 (R). If
statements
years.
years. The
The paper by Iwasaki and Hara (2005)”Generalized are pair (A, B)
equivalent: is controllable, then the following statements
KYP
KYP The paper
years. lemma:
lemma: paper by
by Iwasaki
Unified
Unified frequency
Iwasaki
frequency
and Hara
Hara (2005)”Generalized
and domain
domain inequalities
(2005)”Generalized
inequalities with
with
aare
are
pair (A, B) is controllable, then the following statements
equivalent:
equivalent:
KYP
design lemma:
KYP lemma: Unified
applications”
Unified gotfrequency
got the
frequency domain
IEEE
domain inequalities
Transactions
inequalitiesAuto- with
Auto- are equivalent:
with 1. The inequality
design
design applications”
applications” got the IEEE Transactions 1. The
The inequality  ∗  
matic
design Control
matic Control best
applications”
best got the
paper
paper award
the
award
IEEE
IEEE in
in
Transactions
2006. The
Transactions
2006. The paper
paper
Auto-
Auto- by
by 1. 1. The inequality
inequality  x ∗ x
matic
Tanaka Control
and
matic Control best
Langbort paper
best paper award
(2011)”Symmetric in 2006.
award in 2006. formulation The paper
The paper by by
of 
x 
∗∗ G  x 
≥0 (1)
Tanaka
Tanaka and
and Langbort
Langbort (2011)”Symmetric
(2011)”Symmetric formulation
formulation of
of x
u
x G x
u
x ≥ 00 (1)
Tanaka and Langbort (2011)”Symmetric formulation of u G
G u ≥
≥ 0 (1)
(1)
 Supported by SPbSU, grant 6.38.230.2015. u u
uR, x ∈ Cun×1 , u ∈ C such that
n×1
 Supported by SPbSU, grant 6.38.230.2015. is
is satisfied
satisfied forfor all
all ωω ∈
∈ R, x∈ Cn×1 ,u∈ C such that
∈ R, ∈C ∈C

 Supported
Supported by by SPbSU,
SPbSU, grant
grant 6.38.230.2015.
6.38.230.2015. is
is satisfied
satisfied forfor allall ωω∈ R, x x∈ Cn×1 ,, u u∈ C such
such that
that
Copyright
2405-8963 © IFAC 2015 242 Hosting by Elsevier Ltd. All rights reserved.
Copyright © 2015, IFAC (International Federation of Automatic Control)
IFAC 2015 242
Copyright
Peer review©
Copyright IFAC
©under 2015
IFAC responsibility
2015 242
of International Federation of Automatic
242Control.
10.1016/j.ifacol.2015.09.190
 MICNON 2015
June 24-26, 2015. Saint Petersburg, Russia
Sergei V. Gusev et al. / IFAC-PapersOnLine 48-11 (2015) 238–241 239
iωx = Ax + Bu.
2. There exists a matrix H ∈ SMn (R) satisfying the linear
matrix inequality
 
HA + A∗ H HB
G≥ . (3)
B∗H 0
3. There exist matrices H ∈ SMn (R), h ∈ Rn×1 satisfying
the equation
 
HA + A∗ H HB
G= + hh∗ . (4)
B∗H 0
Here, R (C) is the field of real (complex) numbers; Rm×n
(Cm×n ) is the space of real (complex) m × n matrices;
SMn (R) is the space of symmetrical real matrices of size
n; i is the imaginary unit, and the operation ∗ is the
transposition in the real case and conjunction in the
complex one.
Statement 1 is called the frequency condition, Eq. (4) is
the Lur’e equation. Equivalence of 1 and 2 was proved by
Yakubovich (1962), equivalence of 1 and 3, by Kalman
(1963). For a detailed history of origination and proof
of the lemma the readers are referred to Gusev and
Likhtarnikov (2006).
The Kalman—Yakubovich—Popov lemma is closely con-
nected with the question of solution existence for the
algebraic Riccati equation and inequality. Let
 us represent
Gxx Gxu
the matrix G in the block form as G =
Gux Guu
where Gxx ∈ SMn (R), Gxu ∈ Rn×1 , Gux ∈ R1×n , and
Guu ∈ R. If Guu > 0, then the statements 2 and 3 are
equivalent respectively to the following statements:
2 . There exists a matrix H ∈ SMn (R) satisfying the
Riccati inequality
Gxx − HA − A∗ H − (Gxu − HB)G−1 ∗
uu (Gxu − HB) ≥ 0.
(5)
3 . There exists a matrix H ∈ SMn (R) satisfying the
algebraic Riccati equation
Gxx − HA − A∗ H − (Gxu − HB)G−1 ∗
uu (Gxu − HB) = 0.
(6)
The aim of the paper is to explore the mathematical
foundations of KYP lemma. To this end let us replace the
field of real numbers R in the formulation of lemma by an
abstract ordered field F.
Kalman (1970) had shown that the KYP lemma holds for
any real closed field F. This is very tough condition on
the field. The question arises: can the KYP Lemma be
fulfilled under less restrictive assumption on the field F?
It can be shown that the answer is negative for a standard
formulation of KYP lemma which includes equivalence of
three statements. But we shall show that the equivalence
of first two statements of KYP lemma holds for many other
fields. Because the equivalence of two first statements
of KYP lemma was proved by V.A.Yakubovich we shall
further refer to this part of the KYP lemma as Yakubovich
lemma.
To illustrate the above claims let us formulate Yakubovich
lemma for the field of rational numbers Q.
(2) Theorem 2. Let A ∈ Qn×n , B ∈ Qn×1 and G ∈
SMn+1 (Q). If the pair (A, B) is controllable, then the
following statements are equivalent:
1. The inequality (1) is fulfilled for all ω ∈ Q, x ∈ Qn×1 ,
u ∈ Qc satisfying (2).
c
2. There exists a matrix H ∈ SMn (Q) satisfying (3).
Here Qc is a complex extension of the field Q.
It should be noted that in this case the third statement of
the KYP lemma is not equivalentto firsttwo statements.
−1 0
Consider an example. Let G = , A = 1, B =
0 2
1. Then the statements of the KYP lemma (formulated
in terms of Riccati inequality and equation) take the
following form:
1. 2 − |iω − 1|−2 ≥ 0 ∀ω ∈ Q.
2 . H 2 + 4H + 2 ≤ 0
3. H 2 + 4H + 2 = 0
In case when the field of real numbers is considered all
three statements hold. But if the solution H has to belong
to the field of rational numbers, first two statements hold,
but third does not hold, since the solutions
√ of the Riccati
equation are irrational H1,2 = −2 ± 2.
The proof of the theorem will be given in the next section.
3. YAKUBOVICH LEMMA IN ORDERED FIELDS
, Let F be ordered field. Consider in the field F a new
partial order , defining the totally non-negative
 elements
as sums of squares, i.e. a0 ⇐⇒ a = c2j . This
j
order induces the partial order in the space of symmetric
matrices SMn (F) : A0, if ∀x ∈ Fn×1 x∗ Ax0.
Yakubovich statement.
For all n, A ∈ Fn×n , B ∈ F n×1 and G ∈ SMn+1 (F), if the
pair (A, B) is controllable, then the following statements
are equivalent:
1. For all x ∈ Fnc , u ∈ Fc , ω ∈ F such that iωx = Ax + Bu
the inequality
 ∗  
x x
G 0
u u
holds.
2. There exists H ∈ SMn (F) satisfying the LMI
HA + A∗ H HB
G .
B∗H 0
Here Fc is a complex extension of the field F.
Our purpose is to define for what fields the Yakubovich
statement holds.
Definition 1. The ordered field F has sums of squared
polynomials property (SOSP for short), if every polyno-
mial in one variable p ∈ F[λ], that satisfies the inequality
p(λ)0 for all λ ∈ F, can be represented as sum of squares
of polynomials.
Theorem 3. Yakubovich statement is fulfilled in the field
if and only if this field has SOSP property.
The proof of theorem is given in appendix.
243
MICNON 2015
240 Sergei V. Gusev et al. / IFAC-PapersOnLine 48-11 (2015) 238–241
June 24-26, 2015. Saint Petersburg, Russia
The proof of theorem 2. The theorem 2 is a corollary
of theorem 3. To see this let us note that due to Lagrange
celebrated theorem any positive rational number is a sum
of four squares of rational numbers. This implies that in
the field of rational numbers the partial order defined
in this section coincides with the unique linear order in
this field. In addition Pourchet (1971) had shown that
any non-negative polynomial in one variable with rational
coefficients is a sum of five squared polynomials with
rational coefficients. Thus, the field of rational numbers
has SOSP property.
4. HILBERT’S 17TH PROBLEM AND SOSP
PROPERTY OF THE FIELD OF RATIONAL
FRACTIONS.
One more example of the field that posses SOSP property
is closely connected with Hilbert’s 17th problem.
Hilbert’s 17th problem.
Given a multivariate polynomial with real coefficients that
takes only non-negative values over the reals, can it be
represented as a sum of squares of rational functions?
The positive answer on this question was given by Artin
in 1927.
Consider a field F of rational fractions in k variables,
F = R(µ), µ = (µ1 , . . . , µk ). One can get the following
corollary of Artin’s result:
Corollary . The field of rational fractions has SOSP prop-
erty.
Proof. Take a polynomial in one variable p(λ) ∈ F [λ] with
coefficients from the field F. Suppose that for any f ∈ F,
there exist fractions qj ∈ F such that

p(f ) = qj2 , (7)
j
where qj depends of f. We should prove that in this case
there exist polynomials pj ∈ F[λ] such that
p= p2j . (8)
j
The polynomial p can be considered as a rational fraction
p(λ, µ) in variables λ and µ. From (7) it follows that for all
λ ∈ R and µ ∈ Rk p(λ, µ) ≥ 0. Then from Artin’s result it
follows that 
p= rj2 , (9)
j as a sum of squares of rational functions with coefficients
where rj are rational functions in λ and µ. To complete
the prove we need the following result:
Theorem 4. (Artin—Cassels—Pfister). If the polynomial
in one variable is the sum of squares of rational functions,
then it is sum of squares of polynomials.
The rational fractions rj can be considered as rational
fractions in variable λ with coefficients from the field F.
Applying the theorem 4 to (9) we get (8).
Now, using theorem 3, we can conclude that Yakubovich
statement holds when the field F is a field of rational
fractions. In fact, this result can be strengthened.
244
Theorem 5. Let t = (t1 , . . . , tm ), A ∈ R(t)n×n , B ∈
R(t)n×1 and G ∈ SMn+1 (R(t)). We consider the domain
Θ ⊂ Rm consisting of the points θ = (θ1 , . . . , θm ) for which
the functions A(θ), B(θ), and G(θ) are defined. If the pair
(A(θ), B(θ)) is controllable for all θ ∈ Θ, then the following
statements are equivalent:
1. The inequality
 ∗  
x x
G(θ) ≥0
u u
is satisfied for all θ ∈ Θ, ω ∈ R, x ∈ Cn×1 , and u ∈ C such
that
iωx = A(θ)x + B(θ)u.
2. There exists the matrix H ∈ SMn (R(t)) satisfying the
linear matrix inequality
 
H(θ)A(θ) + A(θ)∗ H(θ) H(θ)B(θ)
G(θ) ≥
B(θ)∗ H(θ) 0
for all θ ∈ Θ for which the matrix H(θ) is defined.
The essence of Theorem 5 can be putted in a nutshell
as follows: if the matrices A, B, and G in Yakubovich
lemma depend rationally on the parameters, then the
linear matrix inequality has solution H which is also a
rational function of the parameters. The proof of theorem 5
can be found in Gusev (2014).
5. EQUIVALENCE OF YAKUBOVICH STATEMENT
AND STATEMENT OF HILBERT’S 17TH PROBLEM.
Hilbert formulated his problem for the field of reals. From
the Artin’s result it follows that the statement of Hilbert’s
17th problem holds for any real closed field. Artin also
proved that the statement holds for the field of rational
numbers.
The original formulation of the statement of Hilbert’s 17th
problem cannot hold in a field where positive elements
are not necessary sums of squares. But it can be shown
that accordingly generalized formulation holds in many
such fields. This generalization of Hilbert’s 17th problem
is following:
The generalized formulation of Hilbert’s 17th
problem.
Given a multivariable polynomial with coefficients from
some ordered field F, all values of which over the field F
are sums of squares. Can this polynomial be represented
from F?
It is clear that this formulation of Hilbert’s 17th problem
coincides with original one for the field of reals and for
any field where any positive element is sum of squares.
The presented generalized formulation of Hilbert’s 17th
problem is a special case of problem that was formulated
and investigated by Zeng (1988, 1998). From corollary of
theorem 3 in Zeng (1998) it follows that the generalized
statement of Hilbert’s 17th problem holds in an ordered
field if and only if the field has SOSP property.
Summing up above mentioned result and the Theorem 3
we get the following statement.
 MICNON 2015
June 24-26, 2015. Saint Petersburg, Russia
Sergei V. Gusev et al. / IFAC-PapersOnLine 48-11 (2015) 238–241 241


Theorem 6. The generalized Yakubovich statement and
 g2(n+1−i) ,
 if i = j,
the generalized statement of Hilbert’s 17th problem are 1
Gi,j = g2n+2−i−j , if |i − j| = 1,
equivalent. 
2
0 otherwise
REFERENCES Taking into account that any vector z = (z1 , z2 , . . . , zn+1 )∗ ,
Gelig, A.K., Leonov, G.A., and Yakubovich, V.A. (1978). satisfying (A.1), has the form z = zn+1 (λn , λn−1 , . . . , 1)∗ ,
Stability of nonlinear systems with nonunique equilib- we get inequality
rium state. Fizmatgiz, Moscow. z ∗ Gz = zn+1
2
g(λ)0 (A.3)
Gusev, S.V. (2014). Kalman–Popov–Yakubovich lemma
for all λ, z, satisfying (A.1). Thus, there exists H ∈
for ordered fields. Automation and Remote Control, 75,
18–33. HMn (Fc ) such that (A.2) is fulfilled. Let G̃ = Re(G +
Gusev, S.V. and Likhtarnikov, A.L. (2006). Kalman– i(M ∗ HN − N ∗ HM )). Due to (A.2) G̃0 and, conse-
Popov–Yakubovich lemma and the S-procedure: A his- quently, there exist vectors hj ∈ F(n+1)×1 such that
torical essay. Automation and Remote Control, 67, 77– 
G̃ = hj h∗j . (A.4)
121.
j
Iwasaki, T. and Hara, S. (2005). Generalized KYP lemma:
Unified frequency domain inequalities with design appli- From the equality in (A.3) it follows that g(λ) =
cations. IEEE Trans. Automat. Control, 50, 41–59. , 1)∗ . After substitution G̃
(λn , λn−1 , . . . , 1)G̃(λn , λn−1 , . . . 
2
Kalman, R. (1963). Lyapunov functions for the problem of from (A.4) we obtain g(λ) = j fj (λ), where fj (λ) =
Lur’e in automatic control. Proc. Nat. Acad. Sci. USA, n
(λ , λ n−1
, . . . , 1)hj . This proves that the field F has SOSP
49, 201–205. property.
Kalman, R. (1970). New algebraic methods in the stability
theory. Proc. 5th Int. Conf. Nonlinear Oscillations, 2,
189–199.
Pourchet, Y. (1971). Sur la representation en somme de
carres des polynomes a une indeterminantee sur un corps
de nombres algbriques. Acta Arithm., 19, 89–104.
Tanaka, T. and Langbort, C. (2011). Symmetric formula-
tion of the Kalman-Yakubovich-Popov lemma and exact
losslessness condition. Proc. 50th IEEE CDC, 5645–
5652.
Yakubovich, V.A. (1962). Solution of some matrix inequal-
ities appearing in the automatic control theory. Dokl.
Acad. Nauk SSSR, 143, 1304–1307.
Zeng, G. (1988). A characterization of preodered fields
with the weak Hilbert property. Proc. American Math-
matical Society, 104, 335–342.
Zeng, G. (1998). A problem about the weak Hilbert
property. Acta Mathematica Sinica, New Series, 14,
481–486.

Appendix A. THE PROOF OF THEOREM 3

The sufficiency part of the theorem is proved in Gusev

(2014). Let us prove that the field F has SOSP property
when Yakubovich statement holds. If Yakubovich state-
ment is fulfilled, then applying lemma 1 from Gusev (2014)
one can see that for all G ∈ SMn+1 (F) the following
statements are equivalent:
1̃. For all z ∈ F(n+1)×1 and λ ∈ F such that
(λN − M )z = 0 (A.1)
the inequality
z ∗ Gz0
holds.
2̃. There exists H ∈ HMn (Fc ) such that
G + i(M ∗ HN − N ∗ HM )0. (A.2)
Here M = (In , 0n×1 ), N = (0n×1 , In ), In is the identity
matrix of dimension n, 0n×1 is the zero matrix of size
n × 1, HMn is a space of Hermitian matrices.
Take a polynomial g(λ) = g2n λ2n + g2n−1 λ2n−1 + . . . +
g0 such that g(λ)0 for all λ ∈ F. Define a matrix
G ∈ SMn+1 (F) as follows:

245