Numerical Approximation Of Generalised Riccati Equations By Iterative Decomposition Method

Document Sample
Numerical Approximation Of Generalised Riccati Equations By Iterative Decomposition Method Powered By Docstoc
					                                                     (IJCSIS) International Journal of Computer Science and Information Security,
                                                     Vol. 9, No. 6, June 2011




           Numerical Approxima
           N         A                            Riccati Equ
                             ation of Generalised R         uations by Iterative
                                Decomp position Meethod
                                                 Kadiri, K.O.2, Odetunde, O.S.3
                                  Adekun Y.A.1*, K
                                       nle,
    1
           artment of Com
        Depa                         e           atics, Babcock University, Ilis
                        mputer Science and Mathema                                        un           ria
                                                                               san-Remo Ogu State, Niger
            dekunleya@g
           ad         gmail.com
    2
            partment Electr
        . Dep                            cs                         technic, Offa, K
                          rical/Electronic Engineering, Federal Polyt                           Nigeria
                                                                                   Kwara State, N
            adiritoyin2007@yahoo.c
           ka                    com
    3
            partment of Ma
        . Dep                          ience, Olabisi O
                         athematical Sci                          versity, Ago-Iw
                                                      Onabanjo Univ                          ate,Nigeria
                                                                                woye, Ogun Sta
            odetunde@ya
           to         ahoo.co.uk


                            I              position Method
Abstract-In this paper, the Iterative Decomp             d          appli in [7] to s
                                                                        ied         solve one-dime            rmonic
                                                                                                 ensional Bihar
           to            neralized Riccat equations. We
is applied t solve the gen              ti            e                 tions. In [8] t method w applied to solve
                                                                    equat             the      was
                                                          s
considered equations with variable coefficients, as well as             ational problem while in [9 delay differ
                                                                    varia             ms          9]           rential
                       fficients. The p
those with constant coeff                          d
                                      present method
                                                                        tion were appr
                                                                    equat                         ng         method.
                                                                                     roximated usin the same m
                          ly                          t
presents solutions as easil computable, fast convergent
                                                                         l             ers,                      ws
                                                                    In all of these pape the results obtained show very
            ies, requiring no discretization Examples are
infinite seri              n               n.           e
                                                                         ble
                                                                    reliab      accuracy     when        mpared
                                                                                                       com            with     exact
           o                            efficiency of the
presented to establish the accuracy and e               e
method.                                                                 tions, and with results obtaine using other k
                                                                    solut                             ed            known
keyword: G
         Generalised Riccati Equations, I
                                        Iterative                      hods.
                                                                    meth
         tion Method.
Decomposit                                                                  The aim of this paper is to continu the
                                                                                                              ue
                                                                        ication of the It
                                                                    appli                            mposition on G
                                                                                        terative Decom            General
                    I.   INTRODUCTION
                          N
                                                                        ati
                                                                    Ricca Equations.
        ccati equations find applicatio in Random
  The Ric                             ons       m
                                                                        layout of the re of the paper is as follows:
                                                                    The l              est          r
                                               d
processes, optimal control, network analysis and
                                                                    In    section   2,     the
                                                                                           t      analysis    of    the      erative
                                                                                                                           Ite
          problem, as well as in Finance mathematics.
diffusion p                            e
                                                                       omposition Me
                                                                    Deco                                      ions is
                                                                                   ethod IDM for Riccati equati
          ons          e            e             d
The solutio of Riccati equations have been obtained
                                                                        ented. We nex present the application o the
                                                                    prese           xt          e             of
         e              umerical metho such as the
using some traditional nu            ods         e
                                                                       hod                                  ples
                                                                    meth developed in section 2 to some examp in
         hod, and the Runge-Kutta m
Euler meth            R                       r
                                  method. Newer
                                                                         on         lusion is drawn on the meth
                                                                    sectio 3. A concl             n           hod, in
                     e            d           e
numerical schemes have been applied by [3]. The
                                                                         on
                                                                    sectio 4.
         domian Decom
popular Ad                       hod     s
                    mposition Meth ADM has
         ied          o             nlinear Riccati
been appli in [2, 6] to solve the non             i
equations. In [10],the Homotopy An
                       H         nalysis Method
                                              d                     II. So
                                                                         olution Techni
                                                                                      ique By Iterat Decompos
                                                                                                   tive     sition

                                    tion of Riccati
was used to obtain the analytic solut             i                    hod
                                                                    Meth
equations.                                                             nsider
                                                                     Con             the         General        Riccati        Diffe

            he                          work is the need
           Th basic motivation for this w              d
                                                                         al
                                                                    rentia Equation.
          tion technique which can be applied with
for a solut            e            e            h
          ase, and requi
relative ea                                     l
                       iring minimal mathematical
           thout any loss of accuracy or e
details, wit              o                            e
                                         efficiency. The            Wher re                                               are scalar
          Decomposition Method (ID
Iterative D           n          DM) has been
                                            n                           tions
                                                                    funct
                                                                    Equaation (i) can b written in the form
                                                                                      be



                                                              310                                 http://sites.google.com/site/ijcsis/
                                                                                                  ISSN 1947-5500
                                                   (IJCSIS) International Journal of Computer Science and Information Security,
                                                   Vol. 9, No. 6, June 2011




        he
 Where th differential operator L is gi
                       o              iven by                     III   Numerical Examples
                                                                        Example 3.1
                                                                  Conssider the R Riccati equation             with      onstant
                                                                                                                        co
The Unive erse operator is thus the int
                        i             tegral operator
                                                    r                 ficients
                                                                  coeff
         y
defined by

                                                                      exact solution of the problem is
                                                                  The e                           m
         the         perator (4) on (2), it follows
Applying t universe op                            s
that


                                                                      Iterative Decom
                                                                  The I                         thod gives
                                                                                    mposition Met



The iterativ Decomposit
            ve             tion method asssumes that thee
unknown f  function                      ssed in terms of
                             can be expres              f
an infinite series of the fo
                           orm

where                                               e
                                          so that the
component   t         b
                 can be determined iteratively
To convey the idea of the method, a well as for
           y                       as           r
                                                                     thus have an ap
                                                                  We t             pproximation
            ess      hnique [7] we c see that (8)
completene of the tech             can          )
           orm
is of the fo

where k is a constant and
                        d           nonlinear term.
                             is the n
We can decompose the nonlinear term a
                        n            as                               e
                                                                  Table 1
                                                                   t      Exact Solution          Approximate             rror
                                                                                                                         Er
                                                                                                solution b IDM
                                                                                                         by
                                                                  0.0
                                                                  0.1    0.099667994                   81
                                                                                              0.09966798                   -8
                                                                                                                      1.3 E-
                                                                  0.2    0.197375320                   04
                                                                                              0.19737370              1.616 E-6
From (8) a (10), equati (9) is equiv
         and          ion          valent to                      0.3    0.291312612                   272
                                                                                              0.29128522              2.738 E-5
                                                                  0.4    0.379948962                   26
                                                                                              0.37974602              2.029 E-4
                                                                  0.5    0.462117157                   031
                                                                                              0.46116290              9.543 E-4
                                                                  0.6    0.537049366                   643
                                                                                              0.53368786              3.362 E-3
                                                                  0.7    0.604367777                   936
                                                                                              0.59467219              9.696 E-3
                                                                  0.8    0.664036770                   089
                                                                                              0.63989270              2.414 E-2
                                                                  0.9    0.716297870                   771
                                                                                              0.66254907              5.375 E-2
                                                                  1.0    0.761594155                   824
                                                                                              0.65196588              1.096 E-1

                                                                                       erical comparison o results for Exam 3.1
                                                                  Table 1 shows the nume                 of               mple
                                                                       ned
                                                                  obtain from our present scheme and the exact solutio at aon
                                                                  numbe of points in the interval [0, 1]. Error is def
                                                                        er                                                fined as



          ,…                                                         mple
                                                                  Exam 3.2
Thus,                                                                sider the Riccat equation with variable coef
                                                                  Cons              ti            h             fficient



                                                                      exact solution of the problem is
                                                                  The e                           m
and
                                                                  Appl              ative Decompo
                                                                      lying the Itera                       od,
                                                                                                osition Metho we
                                                                  have



                                                            311                                http://sites.google.com/site/ijcsis/
                                                                                               ISSN 1947-5500
                                                         (IJCSIS) International Journal of Computer Science and Information Security,
                                                         Vol. 9, No. 6, June 2011




                                                                             quation. Int. M
                                                                            Eq                            6):
                                                                                           Math. For. 2 (56 2759-2770
                                                                             007
                                                                            20
Then,                                                                    [4] H. He, Varia                n
                                                                                         ational iteration method for
                                                                             utonomous ord
                                                                            au                           tial
                                                                                         dinary different systems. Appl.
                                                                            Math. Comput. 1
                                                                            M             114:115-123. 2 2000

                                                                         [5] L. Junfeng, Va                ation method fo
                                                                                            ariational itera              or
                                                                             solving two-poi boundary v
                                                                             s               int           value problems. J.
                                                                             Comput. Appl. 2006
                                                                             C
                                                                             T.Ramesh Rao, The use of Ad
                                                                         [6] T                             domian
                                                                            D
                                                                            Decomposition M  Method for Sol lving Generalissed
We thus ha an approximation for y (t as
         ave                       t)                                       Riiccati Differential Equations. Proc. 6th. IM
                                                                                                            .,           MT-GT
                                                                              onf.           t.
                                                                            Co Math. Stat Appl.: 935-9 2010941.
                                                                             O.
                                                                         [7] O Taiwo, S, Od                Y.
                                                                                             detunde, and Y Adekunle,
Table 2:                                                                     N
                                                                             Numerical App  proximation of One Dimensio  onal
t      Exact Solution    Appproximate            Error                       Biharmonic Eq
                                                                             B              quations by an IIterative
                         Soluution by IDM                                    decomposition Method, Int. J. Math. Sc. 20 (1):
                                                                             d
0.0     0.0              0.0                     0.0                         3
                                                                             34-44. 2009
0.1     0.1                 0000000000
                         0.10                    1.000E-26
                                                                             O.              S.
                                                                         [8] O Taiwo, and S Odetunde, N    Numerical
0.2     0.2                 0000000000
                         0.20                    1.000E-26
0.3     0.3              0.30
                            0000000000           1.257E-26                   A                              l
                                                                             Approximation of Variational Problems by a    an
0.4     0.4              0.40
                            0000000000           3.954E-22                   I
                                                                             Iterative Decommposition Meth  hod, Maejo J. M
                                                                                                                          Math.
0.5     0.5                 0000000000
                         0.50                    1.218E-18                                  433
                                                                             3 (03): 426-4 2009
0.6     0.6                 0000000000
                         0.60                    8.636E-16                   O.              S.
                                                                         [9] O Taiwo, and S Odetunde, Ap   Approximation of
0.7     0.7                 0000000000
                         0.70                    2.220E-13                   D               tial
                                                                             Delay Different Equations by a Decompo       osition
0.8     0.8                 0000000000
                         0.80                    2.717E-11
0.9     0.9                 0000000001
                         0.90                    1.886E-9
                                                                             M                              3:
                                                                             Method, Asian J. Math. Stat. 3 1-7. 2010
1.0     1.0                 0000000837
                         1.00                    8.372E-8
                                                                         [10] Y. Tan, and S. Abbasbandy, Homotopy ana   alysis
            ws                                                 2
Table 2 show the comparison of the exact solution of Example 3.2             m
                                                                             method for quaadratic Riccati differential
             tion obtained by th Iterative Decom
with the solut                 he              mposition Method.
                                                                             equations. Com
                                                                             e             mmun. Nonlin. Sci. Numer. Simul.
                                                                             d             cnsns.2006
                                                                             doi:10.1016/j.c
IV. Conclusion
     From the given exam               revious section,
                         mples in the pr
           t             cheme is very e
we see that our current sc                             e
                                       efficient for the
          f
classes of Riccati equa   ations. The aapproximations  s
          are            y
obtained a reasonably accurate, wh      hen compared   d
          xact           e              ew
with the ex solutions, even for very fe terms of the   e
approxima ating infinite series. Th    he level of     f
                         on
mathematical computatio required is very minimal,
with extremmely low com                 st.
                        mputational cos The present    t
           n              s            nt
method, on the whole, is very efficien and accurate    e
for the Ri               ns            ant
          iccati equation with consta coefficients     s
           with          c
and those w variable coefficients.

REFEREN  NCES
[1] G. Doomian, Solving Frontier problems in
         cs:
    Physic The Decomp                od.
                        position Metho Kluwer
          mic          s,
    Academ Publishers Dordrecht. 19  994

         hnasawi, M. El-
[2] A. Bah               -Tawil, and A. Abdel-Naby,
                                       ns
   Solving Riccati Differential Equation using
         an              ion          Appl.
   Adomia Decompositi Method., Ap Math.
          t.             4
   Comput 157: 503-514 2004

[3] B. Ba               ani,
         atiha, M. Noora I. Hashim, Application of   f
          onal Iteration Method to a Ge
   Variatio              M            eneral Riccati




                                                                   312                              http://sites.google.com/site/ijcsis/
                                                                                                    ISSN 1947-5500