where C] , C2 are the finite upper bounds for ilF(x) l[,. Necessity. Denoting by %i the i-th component .... IV San Diego Symposium in Nonli- near Estimation and Its ...
AN I N T R O D U C T I O N BOUNDED
TO
RATE SYSTEMS
C.Bruni Istituto
- G.Koch di A u t o m a t i c a
Universit~ Via Eudossiana,
di Roma
18 - 00184
Roma
A. Germani CSSCCA - CNR
SUMMARY
In this work which
a new class
the d e n o m i n a t i o n
appears physical
to be quite phenomena
of n o n l i n e a r
"bounded
relevant
systems
rate systems"
for its c a p a b i l i t y
in d i f f e r e n t
is introduced,
is proposed. of m o d e l i n g
field such as biology,
This
for
class
important
ecology,engine-
ering. Bounded trol systems, vestigated
rate systems
situate
so that a bounded
and d e v e l o p e d
between
rate
exploiting
bilinear
and l i n e a r - i n - c o n
system theory may be u s e f u l l y already
available
results.
in-
472
I~ A NEW CLASS
OF N O N L I N E A R
SYSTEMS
The need of i n t r o d u c i n g
nonlinear
cal w o r l d
is p r e s e n t l y
an u n q u e s t i o n e d
important
nontechnical
fields
systems
in m o d e l i n g
fact e s p e c i a l l y
like biology,
ecology,
It is equally well know that the m a i n o b s t a c l e is the d i f f i c u l t i e s linear
classes
situation
classes
by the a n a l y t i c a l
study of a general
no~
motivates
the actual
to spot out specific
which possibly
analytical
study
couple
to the a b i l i t y
step in this d i r e c t i o n
of b i l i n e a r
applicative number
systems
trend,
the advantages of m o d e l i n g
of
relevant
of phenomena.
A noteworth
systems,
Another
theoretical
significant
class
relevant
and optimal
This paper
results
is intended
the i n t r o d u c t i o n quite valid
class
of
the
from an
to achieve
a
[1,2].
of this trend was that
the study of a more
is the systems
are a v a i l a b l e
linear
expecially
in co~
on t h e c o n t r o ~
[1,3,4,5,6]. to give a futher
contribution
a new class of n o n l i n e a r
by x the n - d i m e n s i o n this
results
systems,
control
me line by i n t r o d u c i n g
put vector,
was
in fact appears
example
of n o n l i n e a r
for w h i c h
lability
which
point of v i e w and in the m e a n time a l l o w e d
of i m p o r t a n t
general
usual
socio-economics. in this d i r e c t i o n
presented
of n o n l i n e a r
not too d i f f i c u l t
trol
of
system.
This
class
the physi-
in a number
systems.
along
the sa
Denoting
as
state v e c t o r (t:) and by u the p - d i m e n s i o n
is d e f i n e d
i~
as follows:
o
x(t)
where
the operator
= {(x)
+ Nxu
~:~ ÷ R n, ~ open
cally L i p s c h i t z i a n (2) w i t h at m o s t
11+(x) i!
0
which
is in c o n t r a s t By a s i m i l a r
changes
its
sign
with
proof,
(2.18). we can also
in ~ t h r o u g h
b)
x2 ({) = 0
Recalling
(2.13)
reject
a horizontal
,
the p o s s i b i l i t y tangent
flex
that
x2 ([) < 0
(2.22)
this means:
xl (t) >_ul ([) = s O
This
implies
the e x i s t e n c e
(2.13)
we then
(2.23)
of a 0 < t'
0 , e n t h a l p y v a r i a t i o n p,Cp
, density
and s p e c i f i c
We n o w d e f i n e
in the r e a c t i o n , h e a t of the input
assumed
exothermic
flow.
the c o n s t a n t s :
F = U__U__ (-AH_____~); K4 = ~E Ks = ~ ; Ks V p C p ; K3 = P Cp
(2.48)
and the s t a t e a n d input v a r i a b l e s ;
(2.49)
Ul u =
l
u2
u3
T h e n eq.
(2.47)
=
Ic°
(2.50)
TO
Tk _
take the form:
x = F ( x ) x + Bu
(2.51)
where: ~(KI + K o e - K---i-4 x2 ) F (x) =
0
_ K4 K3 K o e
x~
)I (2.52)
- (Ks + K2
484
-KI B
0
(2.53)
=
,_0
The
KI
function
in the o p e n bounded
0
set
rate
K2 i
F is l o c a l l y
~ = {x@R2
system,
we n o w
S = {x~R2;
for any
choice
itself.
first
equation
in
" It
xl (t o ) e
=
)to(KI+Ko
t
-~ ~t
+I
and
therefore As
> t
-
the
Je
2.6.
e
0}
(2.56)
(which
solution
are of
the o n l y
(2.51)
physi-
stays
(2.51) m a y be i n t e g r a t e d K4 x2 (
imply
xl (t) ~ 0 .
if we a s s u m e
that
there
exists
a time
that:
Other ded
a
o
K1u2 (~) + K 2 U 3
which
bounded range
is a c t u a l l y
K4
(KI + K
xt (t o ) ~ 0
x2 ({) from
(2.54)
(2.55)
e
far as x2
such
o
t
uniformly
that
0 }
any p o s s i b l e
_
xl (t)
with
defined:
in S and of u in U
choices), the
that,
> 0, u2 (t)>0,
of X(to)
Indeed,
show
xl >_ 0, x2>
U = {u : ul (t)
cally meaningful
Lipschitzian
: x2 > 0} . To p r o v e
in
we get:
(2.59)
(t) < 0
with
the
Kinetics
important
examples
are
(2.58)
(2.51)
reactor
systems
< 0
assumed
positiveness
of q u a d r a t i c
the u s u a l l y
adopted
systems
models
of u2 (t) , u3 (t) .
which
are
also
for the n u c l e a r
boun
reactor
kinetics. As
is e x t e n s i v e l y
of a p o i n t region
with
reactor Newton
with
reported one
cooling
[16,17]
group is:
a model
of d e l a y e d
for the
neutrons
free
and one
response feedback
485
(T-T o) + = c =
9 - Ic
= K(v
where
~ is
average Vo'
Co -
the
neutron
reactor I-X
= temperature
density,
To"
= reactor
I/7
= mean
of
fraction
decay
constant
time
c is
with
neutron
heat
- TO )
the
precursor
equilibrium
density,
values
T
is
an
respectively
Furthermore:
generation
I/K
- Vo ) - ? ( T
coefficient
I = precursor 1 = neutron
(2.60)
temperature,
Vo'
8 = delayed
~, + Ic
1
reactivity
time
capacity
for
heat
transfer
to
the
coolant
Defining:
X
=
[xl II X2
=
x3
equations
(2.60)
x
=
(2.61)
CO
_
take
To
the
form:
F(x) x
(2.62)
where: ~B
-
F(x)
The to p r o v e that
=
system is
purpose
the let
S =
~
y
-~
0
K
0
-y
(2.62) uniform us
C~
y
is
y(Vo+Xl)
clearly
the
xl > -
set
Vo,
(2.63)
a quadratic
boundedness
define
{xER3:
-
of
its
one
possible
so
that
all
we
trajectories.
have For
S:
x2 > - C o }
(2.64)
486
and the scalar function V on S:
I
V(X)
= vO
X[
[ F
t . --
0
PROOF.
First
and uniformly
of all,
bounded
(to,Xo) E RIxS~
there
note
that,
being
i n ~ D S, a n d b e i n g exists
a RO =
F,
f locally
u continuous
Lipschitzian
for any
{ ( t , x ) : I t - t o l £a,l[x-XoH ! ~}
such
that:
a) b)
R
C o the R H S
of
t for e a c h Therefore
(1.6)
Furthermore (1.2)
(1.6) ~ in Ro, fixed
is u n i f o r m l y
x and uniformly
admits
t h e R H S of
a local (1.6)
solution
continuous in x.
through
is c o n t i n u o u s
in
(to,Xo).
and recalling
is s u c h that:
l[F(x(t))x(t) + f(x(t)) + N x ( t ) u ( t )
+
{INiJ flu(t)[I
KI (t) ~ K2 (t) a r e
+ Bu(t) [[ t --
RE~£ARK I. A l l systems que
satisfy
solution
REMARK proved of Thm. ses
the p r e v i o u s l y
the a s s u m p t i o n s
3.1.
for
all
solution
of
.
examples
3.1.
and
of b o u n d e d
therefore
admit
rate a uni
in large.
2. A s i m i l a r
in [21 ] . The
in w h i c h
local
o
mentioned
of Thm.
z(t o) E R and any
and both
theorem,
for
set of a s s u m p t i o n s
somehow
more
systems seems
restrictive
are a p p l i c a b l e .
evolving
on group,
to be s i m i l a r
than
the
latter,
is
to that in t h o s e
one ca
490
4. C O N C L U S I O N S
In the previous of b o u n d e d
rate
section we e n l i g h t e n e d
systems.
We showed
processes
may be included
in that class.
From
the t h e o r e t i c a l
point
ded rate
systems
the theory our m i n d
was given
this
should be an important
In p a r t i c u l a r troi would
results
be useful.
b u t i o n was given
in [9]
where
system as a bilinear This
bounded
rate
for b i l i n e a r
systems
are already
[3,4,5,6,21,24]
available
which
a controllability
rive
is concerned,
hopefully
theory
controllability
properties
of the state equation. systems, available
Under
for which [I].
will
for b o u n d e d
some papers
already
exploiting
contr~
for the immune
structure
in a u n i f o r m l y
for the class
We also m e n t i o n
bilinear
stability
the general
a first
bounded
feedback
theoretical already
res-
of a b o u n d e d loop.
results
available
for
results
systems.
As far as c o n t r o l l a b i l i t y sults
and o p t i m a l c o n
is concerned,
of how to achieve
by suitably
and in
task to be pursued.
asymptotic
system
is a good example
result for h o u ~ point out that
on stability, c o n t r o l l a b i l i t y
exploiting
physical
is still to be d e v e l o p e d
As far as s t a b i l i t y
ponse model was p r o v e d rate
3. We must however
of systems
of the class
of important
of v i e w a first general
in Sect.
for such a class
the r e l e v a n c e
how a number
a number
of linear
be a useful rate
[25,26]
starting
re-
systems
point to build
systems. in w h i c h
by looking
it is shown how r o d e
only at the u - d e p e n d e n t p a r t
this aspect b o u n d e d some results
of important
in c o n t r o l
rate
systems
on c o n t r o l l a b i l i t y
behave as
theory
are
491
APPENDIX
In o r d e r
to p r o v e
THEOREM.
Given an open subset
condition
eq.
(I .3)
we m a y
state
following
~ of R n a necessary
~ : ~ + R n to be
for a function
the
and sufficient
locally Lipschitzian
with the
growth property:
II~ ( x ) II ~ C ,
IlxlI + C 2
where CI ,C2 are nonnegative
= F(x)x
Sufficiency.
PROOF.
~
(A.I)
is that there exist
locally Lipschitzian
two functions
and with uniformly
such that:
(x)
bounded
¥x~
constants,
f : ~ + R n which are
F : ~ ÷ R n×n,
bounded range
,
subset
MC~
we
+ f(x)
If
,
(A.2)
VxC
(A.2)
holds,
then
f o r xl , x2
in a n y
closed
have:
11% (xl) - ~ (x2) II= IIF(xJxl
+ f(xl ) - F ( x 2 ) x 2
< liE(x1 )[I Ilxl-x2 tI+IIF(x:)-F(x2)U < maxIIF(xl)II x~ ~ M
- f(x2) II
tlx2 II+[tf(xl )-f(x2)I1
IIx~-x2 II+ maxllx2 II L F x2 E M
II x l - x 2 If+ Lf~xl-x211