ON THE CATEGORY OF ORDERED TOPOLOGICAL MODULES OPTIMIZATION AND LAGRANGE’S PRINCIPLE

IJESM Volume 2, Issue 1 ISSN: 2320-0294
_________________________________________________________
A Quarterly Double-Blind Peer Reviewed Refereed Open Access International e-Journal - Included in the International Serial Directories
Indexed & Listed at: Ulrich's Periodicals Directory ©, U.S.A., Open J-Gage, India as well as in Cabell’s Directories of Publishing Opportunities, U.S.A.
International Journal of Engineering, Science and Mathematics
http://www.ijmra.us
140
March
2013
ON THE CATEGORY OF ORDERED TOPOLOGICAL
MODULES OPTIMIZATION AND LAGRANGE’S
PRINCIPLE
P.A. Otieno
K.C. Sogomo*
M.N. Gichuki*
Abstract
A category is an algebraic structure made up of a collection of objects linked together by
morphisms. As a foundation of mathematics, categories were created as a way of relating
algebraic structures and systems of topological spaces In this paper we define a derivative using
cones in the category of topological modules and use the Lagrange’s principle to obtain
optimization results in the category.
 Department of mathematics, Egerton University, Egerton, Kenya

_________________________________________________________
141
March
2013
1. Introduction
The main problem that seems to make the study of optimization in categories difficult is the fact
that it is algebraic in nature yet most of the optimization is studied in classical analysis. Since the
innovative use of infinitesimals by Lawvere (1963) and Kock (1981) it has been possible to
study some parts of analysis in such toposes.Most of the extrema problems involve the order
properties of the real line and this explains why extrema properties of a complex variable do not
exist.Sukhinin (1982) introduced the idea of extrema in spaces without norm that is applicable
even to functions of complex variables. It is this idea we seek to adopt in the ordered category of
topological modules. The systematic study and use ordered vector spaces and cones in
mathematics begun around the world after 1950 mainly through the efforts of Russian, Japanese,
German and Dutch mathematicians. The notion of cones is important in many areas; two notable
areas are optimization theory and the fixed point theory. Since cones are being employed to solve
optimization problems, the theory of ordered vector spaces is an indispensable tool for solving a
variety of applied problems in diverse areas such as engineering, econometrics and the social
sciences (Charambos and Rabee, 2007).
A subset C of a vector space V is a cone if αx + βy belongs to C for any positive scalars α and β
and any x,y in C.This concept is meaningful for any vector space that allows the concept of
positive scalars such as spaces over rational ,algebraic and even the real numbers. It follows
hence that the empty set, the space V and any linear sub space of V including the trivial sub
space {} are convex cones by this definition. The intersection of two convex cones in the same
vector space is again a convex cone but their union may fail to be one. The family of convex
cones is closed under arbitrary linear maps and particularly if C is a convex cone so is its
opposite C and C C is the largest linear subspace contained in C (Rockafellar, 1997).
If X is within the Euclidean space, the cone on X is homeomorphic to the union of lines from X
to another point. That is, the topological cone agrees with the geometric cone when defined. All
cones are path connected since every point can be connected to the vertex by the homotopy
ht(x,s) = {x,(1-t)s}.In algebraic topology, cones are used precisely because they embed a space as
a subspace of a contractible space (Allen,2002).Since cones can be topologically closed, open or
both then continuity being a central concept in topology can be defined using open
neighbouhoods that is if X and Y are topological spaces , a function f: XY is continuous if

_________________________________________________________
142
March
2013
and only if for all open neighbouhoods of B of f(x) there exists a neighbouhood A of X such
thatA  f -1
(B), (Gaal and Steven, 2009).
2. Literature review
A category is an algebraic structure that comprises ‘objects’ that are linked by ‘arrows’. A
category has two basic properties, the ability to compose the arrows associatively and the
existence of an identity arrow for each object. A simple example is the category of sets whose
objects are sets and whose arrows are functions. On the other hand, a monoid can be understood
as a special sort of category and so can any pre-order. Generally objects and arrows may be
abstract entities of any kind and the notion of category provide a fundamental and abstract way
to describe mathematical entities and their relationships.This is the central idea of category
theory, a branch of mathematic which seeks to generalize all of mathematics in terms of objects
and arrows independent of what the object and arrows represent.
Virtually every branch of modern mathematics can be defined in terms of categories and in doing
so revealing deep insights and similarities between seemingly different areas of mathematics.
Categories were introduced by Eilenberg and Mac Lane (1945).
A topological module is a module over a topological ring such that scalar multiplication and
addition are continuous. In abstract algebra, a module over a ring is a generalized notion of a
vector space, wherein the corresponding scalars lie in an arbitrary ring. An abelian topological
group that is a module over a topological ring R, in which the multiplication mapping R x A→A,
taking (r, α) to rα is required to be continuous. A right topological module is defined
analogously. Every sub module B of a topological module A is a topological module. If the
module A is separated and B is closed in A, then A/B is a separated module (Bourbaki, 1966).
The left R- modules together with their module homomorphisms form a category written as R-
Mod which is an abelian category, (Anderson and Fuller, 1992).
Given a ring R and an R- module, a descending filtration of M is a decreasing sequence of sub
modules Mn.This is therefore a special case of the notion of groups with the additional condition
that the sub groups be sub modules. The associated topology is defined as for groups. An
ascending filtration is defined in the same way (Oksendal and Bent, 2003).In topology and
analysis, filters are used to define convergence in a manner similar to the role of sequences in

_________________________________________________________
143
March
2013
metric spaces. Sequences are usually indexed by natural numbers which are a totally ordered
set.Thus, limits in spaces can be defined using sequences (Victor, 2012).
2.2 Boundary and Extrema
Given two real axes X and Y and a function f: X  Y, that takes a closed bounded set A  X to
a closed bound set B  Y then problems of maxima and minima involves finding points in A that
are mapped by to either the maximum or minimum points of B. If the closed bounded set is not
the real line then there will be no maximum or minimum. It is possible however to introduce a
related idea of extremal point as a point in A that is mapped by the function f to the boundary of
B. This is done using locally convex topological vector spaces with bounded topology. This is
done because classical calculus operations work well up to the abstraction of Banach spaces but
not beyond. Further developments have shown that classical calculus can still work well in
topological vector spaces more general than Banach spaces provided the topology used is
bounded. Another good reason for this topological space is that it is amenable to category
theoretic approach of study (Andreas and Peter, 1997).
2.3 Classical Lagrange Multiplier
One of the most common problems in calculus is that of finding maxima and minima of a
function, but it is often difficult to find a closed form for the function being extremized. Consider
the optimization problem: maximize f(x, y) subject to g(x, y). We introduce a new variable λ
called the Lagrange function defined by:
(x, y, λ) = f(x, y) + λ {g(x, y) – c} Where the λ term may added or subtracted. If f(x, y) is a
maximum for the original constrained problem then  λ such that (x, y, λ) is a stationary point
for the Lagrange function. However, not all stationary points yield a solution to the original
problem. Thus, this method yields a necessary condition for optimality in constrained problems
(Vapnyarskii, 2001). The role of Lagrange multipliers in locally convex spaces and the concept
of extrema can be extended to topological spaces without norm and even to functions of complex
variables.

_________________________________________________________
144
March
2013
2.4 Lagrange Multiplier in locally convex spaces without norm
Sukhinin studied Lagrange multipliers in vector spaces with bounded topology and it is this idea
we intend to generalize to categories of topological modules. To obtain results similar to those
found in classical analysis, Sukhinin (1982) used the idea of cones in linear topological spaces
without norm. We now follow Sukhinin (1982) to get these important definitions;
2.4.1Cone in a Topological Vector Space without Norm
Let X be a linear topological vector space without norm. S  X,xoX, then
S(xo)={hX:UBt(0,):xo+t(h+UB)S} and S+(xo)=X{(XS-(xo)}are called
cones. U is neighbourhood of zero, B is a bounded set and  is a certain system of bounded,
convex sets in U containing zero.
2.4.2 Small Map
Let xoS  X and let Z be a topological vector space. A map r: S → Z will be called β small
relative to the cone K at the point xo if hK,VU,,B :(t,xh+UB,xo+txS),V is a
neighborhood of zero in Z.
2.4.3 Differentiable Map
A map f: S → Z is called β differentiable relative to K at the point xo if
f (xo + h) - f (xo) = f 
(xo) h + r (h). Where f (xo) is a linear continuous operator and the map r: S
 Z is β small relative point xo.
2.4.4 Critical Map
Let S  X and let Z be a linear space. We say that a map f: S  Z is  critical relative to the
cone K at the point xoS if zZ,hK,U,,Bt(0,):tzf{[xo+t(h+UB)]S}f(xo).
2.4.5 Conditional Minimum of Maps
Let xoS  X, let Z be ordered linear space and let f: S Y be a map satisfying
f (xo) = 0.We say that xo is a point of conditional β minimum of the map F: S  Z relative to the
cone K under the condition f (x) = 0 provided
hKUBt(0,):x[xo+t(h+UB) ]S,f(xo) = 0F(x)F(xo).

_________________________________________________________
145
March
2013
2.4.6 Theorem 1: Let xoЅ  X, let K1 be a cone in X with vertex at zero, and suppose that the
map f: Ѕ →Y is such that
hK1,UBt(0,)xt(UB):[xo+ th + xS][f(xo+ th + x) = f(xo)] and
f (xo) = 0.Further, let Z be an ordered topological vector space, suppose that the map F: S → Z is
β differentiable relative to K1 at the point xo.If xo is a point of conditional  minimum of the map
F relative to K1 and under the condition f (xo) = 0 then F(xo)(h)0 for hK1.
Sometimes, the theorem above can be given the form of the rule of Lagrange multipiers, as
shown below;
Let Z be an ordered topological vector space, g be a continuous linear operator from X into Z, let
A be a linear (not necessary continuous) operator from X into Y, K be a cone in X, and suppose
thath (Ker A) K: g (h) ≥0…...*.
Proposition1:Ssuppose (*) above holds, g is a linear is open (AX=Y), and that K = X, then  an
operator   (Y, Z) such that (g+ A) x = 0 for some x X. Indeed to maximize f(x, y)
subject g(x, y), we introduce a new variable  and
 (x, y,) = f(x, y) + [g (x, y) - c].
If f(x, y) is maximum for constrained problem then : (x, y,  ) is a stationary point for (x, y,
z) i.e. (x, y,)  ∂Y. since K = X, K is cone, AX = Y, then Y is a cone. In this case (x, y,) 
vertex of Y which is zero. In our case f(x, y) = g(x, y), [g(x, y) – c] = Ax hence (g + oA) x = 0
implies optimality
of .
3. Preliminaries and Definitions
To come up with a criterion for finding extrema in categories with infinitesimals using a
modification of Sukhinin’s definition of extrema in topological vector space without norms. We
will need the idea of boundedness and boundary in categories. These will be defined in the
required categories. The next step will be to define the extremal object in an ordered category of
topological modules with infinitesimals and finally prove a result on conditions for existence of
extrema. In this regard we intend to formulate and prove theorem 1 in an ordered category of
topological modules.

_________________________________________________________
146
March
2013
3.2 Methodology
The method which we intend to use to get results in optimization in such categories hinges on the
following definitions;
3.2.1 Amazingly tiny object models (a.t.o.m)
Let R be a ring, then D = {d  R: d2
= 0} is called the collection of a.t.o.m. (amazingly tiny
object models in R).A map r: D  D will be called a homomorphism of infinitesimals.
3.2.2 Modules with a.t.o.m
X/D is a factor module with a.t.o.m. if X is a module.
3.2.3 Bounded Module
A module X is bounded if  x  X, d  D, then d.x  D.
3.2.4 Boundary
Let X, Y be topological modules with infinitesimals such that X - Y D, then X-Y is the
boundary of Y and we write  Y.
3.2.5 Extremal Object
Let X be a category of topological modules with infinitesimals, f be morphism in X, xX is an
extremal object of the morphism f(x)  X.
Using the definitions, it should be possible to get a result similar to Lagrange’s method of
multipliers in the category of topological modules by modifying Sukhinin’s method.
3.2.6 Cones in Topological Modules
Let X be a topological module, U a neighbourhood of zero, B a bounded neighbourhood of zero
in X. Then {hX:xo+t(h + UB)  S}where S  X is a cone.
4. Main Results
Theorem 2: Let X and Y be ordered topological modules, x0SX,K1 be a cone in X with
vertex at zero and suppose that the map f: SY satisfies the condition (x0 + dh + xS)  [f(x0 +
dh + x) = f(x0)] and f(x0) = 0. Further, let Z be an ordered topological module, F: S  Z be

_________________________________________________________
147
March
2013
synthetically differentiable relative to K1 at x0. If x0 is a point of conditional minimum of the
map F relative to K1 and under the condition f(x0) = 0, then F(x0) (h)  t (h+UB) for hK1
.......………………(1)
Proof: Assume that x0 = 0, F(0) = 0. Let hK1 and VD in Z. Set r(x) = F(x) - F(0) (x) and set
VD in D satisfying V + VV. Then  U1 and 1 such that F (0) (U1)  V and BdD
such that
(x[d(h + U1B)] S, f(x) = 0. Which implies that [r(x)dV][F(x)  t(h + UB)].
Further  xd (U1B) for which (dh + x) S and f (dh + x) = 0.
Here, F(dh + x) U1B and F (dh + x) C, where C is a positive cone in Z. Further,
F(0)(dh + x) + r(dh + x)  d F (0)(h) + F(0)(x) + r(dh + x). Then,
F(0)(h) + F(0)(d-1
x) + d-1
r(dh + x) [F(0)(h) + V + V] C[F(0)(h) + V]  C, that is
[F(0)(h) + V] C  , since C is closed and V is arbitrary, F(0)(h) C, that is F (0)(h)  d(h +
UB).
Sometimes, the condition (1) can be given the form of the rule of Lagrange multipliers.
Let Z be an ordered topological module, g be a continuous linear operator from X into Z, A be a
linear operator from X into Y, K be cone in X and suppose that;
h(Ker A) K: g(h)d(h + UB)……………………..……………(2)
Proposition 2: Suppose that (2) holds, g is linear, A is open (AX = Y), and that A = X then there
exists on operator
 (Y, Z) such that (g + oA) = 0 for some xX. Where (Y, Z) is a class of continuous
linear maps from Y into Z.
Proof:
From (2) it follows that kerA  Ker g. Now we set y = -gA-1
for some y  Y. The continuity
of  is a consequence of the continuity of g and the openness of A.
Proposition 3
If f is a linear and continuous function, then it maps a cone to a cone.

_________________________________________________________
148
March
2013
Proof: {d (h + UB)} is a cone since f (x0 + d (h + UB)} - f(x0) = f(x0){d(h + UB)} + r(h +
UB),
And since the function f is linear, then it follows that;
f (x0 + d (h + UB)} – f(x0) = d {f(x0) (h + UB)} + r (h + UB), since the left hand side is a
cone, then it follows that the right hand side is also a cone with vertex at zero. Hence the
function f maps a cone to a cone.
5. Conclusion
The results obtained in this work is a great contribution to the study of optimization in the
category of ordered topological modules
6. References
[1]. Allen, H., (2002).Algebraic Topology, Cambridge University Press, Cambridge.
[2]. Anderson, F.W and Fuller, K.R., (1992).Graduate Texts in mathematics, (second ed.)
Springer-Verlag.
[3]. Andreas, K and Peter, W.M., (1997).The convenient setting of Global Analysis,
American Mathematical Society, U.S.A.
[4]. Bourbaki, N., (1996). Elements of Mathematics, General topology, Addison-Wesley
Education Publishers, Boston.
[5]. Charambos, D.A and Rabee, T., (2007).Cones and duality, American Mathematical
Society, U.S.A.
[6]. Kock, A., (1981). Synthetic Differential Geometry. Cambridge University Press,
Cambridge.
[7]. Lawvere, F.W., (1963). Sematics of Algebraic Theories, Ph. D Thesis. Columbia
University, Columbia.
[8]. Mac Lane, S and Eilenberg, S., (1945). Transactions of the American Mathematical
Society, No.2 pp.231-294.
[9]. Oksendal and Bernt, K., (2003).Stochastic Differential Equations, an Introduction
with applications, Springer – Verlag.
[10]. Rockafellar, R.T., (1997).Convex Analysis, Princeton University Press, Princeton.

_________________________________________________________
149
March
2013
[11]. Sukhinin,M.F., (1982).The role of Lagrange Multipliers in locally Convex Spaces,
Siberian Journal of mathematics,No.23 pp.153 – 165.
[12]. Vapnyarskii, I.B., (2001).Lagrange multipliers, Springer – Verlag.
[13]. Victor, P., (2012).Filters on partially ordered sets and generalizations, International
Journal of Pure and Applied Mathematics, No.74 (1) pp.55 – 119.

ON THE CATEGORY OF ORDERED TOPOLOGICAL MODULES OPTIMIZATION AND LAGRANGE’S PRINCIPLE

More Related Content

What's hot (18)

Similar to ON THE CATEGORY OF ORDERED TOPOLOGICAL MODULES OPTIMIZATION AND LAGRANGE’S PRINCIPLE (20)

More from IJESM JOURNAL (20)

Recently uploaded (20)

ON THE CATEGORY OF ORDERED TOPOLOGICAL MODULES OPTIMIZATION AND LAGRANGE’S PRINCIPLE