reduce-historical/r38/packages/odesolve/odesolve.rlg

Tue Feb 10 12:26:51 2004 run on Linux
% Tests and demonstrations for the ODESolve 1+ package --
% an updated version of the original odesolve test file.

% Original Author: M. A. H. MacCallum
% Maintainer: F.J.Wright@Maths.QMW.ac.uk

ODESolve_version;


ODESolve 1.065

on trode, combinelogs;



%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% First-order differential equations
% (using automatic variable and dependence declaration).

% First-order quadrature case:
odesolve(df(y,x) - x^2 - e^x);


*** Dependent var(s) assumed to be y 

*** Independent var assumed to be x 

*** depend y , x
This is a linear ODE of order 1.
It is solved by quadrature.

                       x    3
    3*arbconst(1) + 3*e  + x
{y=---------------------------}
                3


% First-order linear equation, with initial condition y = 1 at x = 0:
odesolve(df(y,x) + y * tan x - sec x, y, x, {x=0, y=1});

This is a linear ODE of order 1.
It is solved by the integrating factor method.

General solution is {y=arbconst(2)*cos(x) + sin(x)}


Applying conditions {{x=0,y=1}}


{y=cos(x) + sin(x)}

odesolve(cos x * df(y,x) + y * sin x - 1, y, x, {x=0, y=1});

This is a linear ODE of order 1.
It is solved by the integrating factor method.

General solution is {y=arbconst(3)*cos(x) + sin(x)}


Applying conditions {{x=0,y=1}}


{y=cos(x) + sin(x)}


% A simple separable case:
odesolve(df(y,x) - y^2, y, x, explicit);

This is a nonlinear ODE of order 1.
It is separable.

Solution before trying to solve for dependent variable is 

arbconst(4)*y - x*y - 1=0


           1
{y=-----------------}
    arbconst(4) - x


% A separable case, in different variables, with the initial condition
% z = 2 at w = 1/2:
odesolve((1-z^2)*w*df(z,w)+(1+w^2)*z, z, w, {w=1/2, z=2});


*** depend z , w
This is a nonlinear ODE of order 1.
It is separable.

                                                   2    2
General solution is {4*arbconst(5) - 2*log(w*z) - w  + z =0}


                         1
Applying conditions {{w=---,z=2}}
                         2


                    2      2
{ - 8*log(w*z) - 4*w  + 4*z  - 15=0}


% Now a homogeneous one:
odesolve(df(y,x) - (x-y)/(x+y), y, x);

This is a nonlinear ODE of order 1.
It is of algebraically homogeneous type 
solved by a change of variables of the form `y = vx'.

                        2            2
{arbconst(6) + sqrt( - x  + 2*x*y + y )=0}


% Reducible to homogeneous:
% (Note this is the previous example with origin shifted.)
odesolve(df(y,x) - (x-y-3)/(x+y-1), y, x);

This is a nonlinear ODE of order 1.
It is quasi-homogeneous if the result of shifting the origin is homogeneous ...
It is of algebraically homogeneous type 
solved by a change of variables of the form `y = vx'.

                        2                  2
{arbconst(7) + sqrt( - x  + 2*x*y + 6*x + y  - 2*y - 7)=0}


% and the special case of reducible to homogeneous:
odesolve(df(y,x) - (2x+3y+1)/(4x+6y+1), y, x);

This is a nonlinear ODE of order 1.

                                   2
It is separable after letting y + ---*x => y
                                   3


{49*arbconst(8) - 3*log(14*x + 21*y + 5) - 21*x + 42*y=0}


% A Bernoulli equation:
odesolve(x*(1-x^2)*df(y,x) + (2x^2 -1)*y - x^3*y^3, y, x);

This is a nonlinear ODE of order 1.
It is of Bernoulli type.

                          5
  1    5*arbconst(9) + 2*x
{----=----------------------}
   2          4      2
  y        5*x  - 5*x


% and finally, in this set, an exact case:
odesolve((2x^3 - 6x*y + 6x*y^2) + (-3x^2 + 6x^2*y - y^3)*df(y,x), y, x);

This is a nonlinear ODE of order 1.
It is exact and is solved by quadrature.

                     4       2  2       2      4
{4*arbconst(10) + 2*x  + 12*x *y  - 12*x *y - y =0}


%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%%

% Now for higher-order linear equations with constant coefficients

% First, examples without driving terms
% A simple one to start:
odesolve(6df(y,x,2) + df(y,x) - 2y, y, x);

This is a linear ODE of order 2.
It has constant coefficients.

     (7*x)/6
    e       *arbconst(12) + arbconst(11)
{y=--------------------------------------}
                   (2*x)/3
                  e


% An example with repeated and complex roots:
odesolve(ode := df(y,x,4) + 2df(y,x,2) + y, y, x);

This is a linear ODE of order 4.
It has constant coefficients.

{y=arbconst(16)*sin(x) + arbconst(15)*cos(x) + arbconst(14)*sin(x)*x

  + arbconst(13)*cos(x)*x}


% A simple right-hand-side using the above example:
odesolve(ode = exp(x), y, x);

This is a linear ODE of order 4.
It has constant coefficients.
Constructing particular integral using `D-operator method'.

{y=(4*arbconst(20)*sin(x) + 4*arbconst(19)*cos(x) + 4*arbconst(18)*sin(x)*x

                                  x
     + 4*arbconst(17)*cos(x)*x + e )/4}


ode := df(y,x,2) + 4df(y,x) + 4y - x*exp(x);


                                x
ode := df(y,x,2) + 4*df(y,x) - e *x + 4*y

% At x=1 let y=0 and df(y,x)=1:
odesolve(ode, y, x, {x=1, y=0, df(y,x)=1});

This is a linear ODE of order 2.
It has constant coefficients.
Constructing particular integral using `D-operator method'.

                                                                 3*x        3*x
                        27*arbconst(22) + 27*arbconst(21)*x + 3*e   *x - 2*e
General solution is {y=---------------------------------------------------------
                                                    2*x
                                                27*e

   }


Applying conditions {{x=1,y=0,df(y,x)=1}}


       3*x        3*x      3        3       2         2
    3*e   *x - 2*e    - 6*e *x + 5*e  + 27*e *x - 27*e
{y=-----------------------------------------------------}
                              2*x
                          27*e


% For simultaneous equations you can use the machine, e.g. as follows:

depend z,x;


ode1 := df(y,x,2) + 5y - 4z + 36cos(7x);


ode1 := 36*cos(7*x) + df(y,x,2) + 5*y - 4*z

ode2 := y + df(z,x,2) - 99cos(7x);


ode2 :=  - 99*cos(7*x) + df(z,x,2) + y

ode := df(ode1,x,2) + 4ode2;


ode :=  - 2160*cos(7*x) + df(y,x,4) + 5*df(y,x,2) + 4*y

y := rhs first odesolve(ode, y, x);

This is a linear ODE of order 4.
It has constant coefficients.
Constructing particular integral using `D-operator method'.

y := arbconst(26)*sin(x) + arbconst(25)*cos(x) + arbconst(24)*sin(2*x)

      + arbconst(23)*cos(2*x) + cos(7*x)

z := rhs first solve(ode1,z);


z := (4*arbconst(26)*sin(x) + 4*arbconst(25)*cos(x) + arbconst(24)*sin(2*x)

       + arbconst(23)*cos(2*x) - 8*cos(7*x))/4

clear ode1, ode2, ode, y, z;


nodepend z,x;



% A "homogeneous" n-th order (Euler) equation:
odesolve(x*df(y,x,2) + df(y, x) + y/x + (log x)^3, y, x);

This is a linear ODE of order 2.
It has non-constant coefficients.
It is of the homogeneous (Euler) type and is reducible to a simpler ODE ...
It has constant coefficients.
Constructing particular integral using `D-operator method'.

                                                                    3
{y=(2*arbconst(28)*sin(log(x)) + 2*arbconst(27)*cos(log(x)) - log(x) *x

               2
     + 3*log(x) *x - 3*log(x)*x)/2}


% The solution here remains symbolic (because neither REDUCE nor Maple
% can evaluate the resulting integral):
odesolve(6df(y,x,2) + df(y,x) - 2y + tan x, y, x);

This is a linear ODE of order 2.
It has constant coefficients.
Constructing particular integral using `D-operator method'.
But cannot evaluate the integrals, so ...
Constructing particular integral using `variation of parameters'.

                    7
The Wronskian is --------
                     x/6
                  6*e


       (7*x)/6                                  (7*x)/6        sin(x)
{y=(7*e       *arbconst(30) + 7*arbconst(29) - e       *int(-------------,x)
                                                              x/2
                                                             e   *cos(x)

             (2*x)/3
            e       *sin(x)          (2*x)/3
     + int(-----------------,x))/(7*e       )}
                cos(x)


end;


Time for test: 510 ms, plus GC time: 40 ms