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

REDUCE Development Version, Wed Sep 13 20:40:41 2000 ...







ODESolve 1.065

% -*- REDUCE -*-
% The Postel/Zimmermann (11/4/96) ODE test examples.
% Equation names from Postel/Zimmermann.
% This version uses REDUCE-style variable notation wherever possible.

on trode;


on div, intstr;

  off allfac;

            % to look prettier

% 1  Single equations without initial conditions
% ==============================================

% 1.1 Linear equations
% ====================

depend y, x;



% (1) Linear Bernoulli 1
odesolve((x^4-x^3)*df(y,x) + 2*x^4*y = x^3/3 + C, y, x);

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

      - 2*x                1     -2    1       1
    e      *arbconst(1) + ---*c*x   + ---*x - ---
                           2           6       4
{y=-----------------------------------------------}
                     2
                    x  - 2*x + 1


% (2) Linear Bernoulli 2
odesolve(-1/2*df(y,x) + y = sin x, y, x);

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

    2*x                2            4
{y=e   *arbconst(2) + ---*cos(x) + ---*sin(x)}
                       5            5


% (3) Linear change of variables (FJW: shifted Euler equation)
odesolve(df(y,x,2)*(a*x+b)^2 + 4df(y,x)*(a*x+b)*a + 2y*a^2 = 0, y, x);

This is a linear ODE of order 2.
It has non-constant coefficients.
It is of the homogeneous (Euler) type (with shifted coefficients) 
and is reducible to a simpler ODE ...
It has constant coefficients.

                 2                                    2
    arbconst(4)*a *x + arbconst(4)*a*b + arbconst(3)*a
{y=-----------------------------------------------------}
                    2  2              2
                   a *x  + 2*a*b*x + b


% (4) Adjoint
odesolve((x^2-x)*df(y,x,2) + (2x^2+4x-3)*df(y,x) + 8x*y = 1, y, x);

This is a linear ODE of order 2.
It has non-constant coefficients.
But ODESolve cannot solve it using linear techniques, so ...
Interchanging dependent and independent variables ...

              2                          3              3            2  2
 - df(x,y,2)*x  + df(x,y,2)*x + 8*df(x,y) *x*y - df(x,y)  + 2*df(x,y) *x

            2              2
 + 4*df(x,y) *x - 3*df(x,y)

This is a nonlinear ODE of order 2.
Interchanging dependent and independent variables ...

           2                            2
df(y,x,2)*x  - df(y,x,2)*x + 2*df(y,x)*x  + 4*df(y,x)*x - 3*df(y,x) + 8*x*y - 1

ODE simplifier loop interrupted! 
ODESolve cannot solve this ODE!

            2                            2
{df(y,x,2)*x  - df(y,x,2)*x + 2*df(y,x)*x  + 4*df(y,x)*x - 3*df(y,x) + 8*x*y - 1

 =0}


% (5) Polynomial solutions
% (FJW: currently very slow, and fails anyway!)
% odesolve((x^2-x)*df(y,x,2) + (1-2x^2)*df(y,x) + (4x-2)*y = 0, y, x);

% (6) Dependent variable missing
odesolve(df(y,x,2) + 2x*df(y,x) = 2x, y, x);

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

Performing trivial order reduction to give the order 1

 linear ODE with coefficients (low -- high): {2*x,1}

It is solved by the integrating factor method.

                                     1
Solution of order-reduced ODE is {{-----},1}
                                      2
                                     x
                                    e


                                                      1
Restoring order, y => df(y,x,1), to give: df(y,x)={{-----},1} and re-solving ...
                                                       2
                                                      x
                                                     e


                  1
{y=arbconst(6) + ---*sqrt(pi)*arbconst(5)*erf(x) + x}
                  2


% (7) Liouvillian solutions
% (FJW: INTEGRATION IMPOSSIBLY SLOW WITHOUT EITHER ALGINT OR NOINT OPTION)
begin scalar !*allfac;  !*allfac := t;  return
   odesolve((x^3/2-x^2)*df(y,x,2) + (2x^2-3x+1)*df(y,x) + (x-1)*y = 0,
      y, x, algint);
end;

This is a linear ODE of order 2.
It has non-constant coefficients.
It is exact, and the following linear ODE of order 1 is a first integral:

         3              2    2
df(y,x)*x  - 2*df(y,x)*x  + x *y - 2*x*y + 2*y=g10
It is solved by the integrating factor method.

                -1
     - 1/2   - x           - 1/2
{y=x      *e      *(x - 2)

                                       1/x
                                      e   *sqrt(x - 2)
 *(arbconst(8) + arbconst(7)*int(--------------------------,x))}
                                           2
                                  sqrt(x)*x  - 2*sqrt(x)*x

% NB: DO NOT RE-EVALUATE RESULT WITHOUT TURNING ON ALGINT OR NOINT SWITCH

% (8) Reduction of order
% (FJW: Attempting to make explicit currently too slow.)
odesolve(df(y,x,2) - 2x*df(y,x) + 2y = 3, y, x);

This is a linear ODE of order 2.
It has non-constant coefficients.
But ODESolve cannot solve it using linear techniques, so ...
Interchanging dependent and independent variables ...

                        3              3            2
 - df(x,y,2) + 2*df(x,y) *y - 3*df(x,y)  - 2*df(x,y) *x

This is a nonlinear ODE of order 2.

                                                                        - 3
This ODE can be simplified by the independent variable shift y => y - ------
                                                                        2

                                  3              2
 to give:  - df(x,y,2) + 2*df(x,y) *y - 2*df(x,y) *x=0


                                                                            G22
This ODE is equidimensional in the independent variable y -- applying y => e

 to transform to the simpler ODE: 

                            3              2
 - df(x,G22,2) + 2*df(x,G22)  - 2*df(x,G22) *x + df(x,G22)=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

                        3        2
 - df(G23,x)*G23 + 2*G23  - 2*G23 *x + G23=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

                    2
 - df(G23,x) + 2*G23  - 2*G23*x + 1=0

This is a nonlinear ODE of order 1.

It is of Riccati type and transforms into the linear second-order ODE: 

df(G23,x,2) + 2*df(G23,x)*x + 2*G23=0

It has non-constant coefficients.
It is exact, and the following linear ODE of order 1 is a first integral:

df(G23,x) + 2*G23*x=g26
It is solved by the integrating factor method.
Restoring order to give these first-order ODEs ...

sqrt(pi)*arbconst(9)*df(x,G22)*erf(i*x)*i - sqrt(pi)*arbconst(9)*erf(i*x)*i*x

     2
    x
 - e  *arbconst(9) - 2*df(x,G22) + 2*x=0

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

df(x,G22)=0

This is a linear ODE of order 1.
It is solved by quadrature.

{arbconst(10) + sqrt(pi)*arbconst(9)

                               erf(i*x)
 *int(-----------------------------------------------------------,x)*i
                                             2
                                            x
       sqrt(pi)*arbconst(9)*erf(i*x)*i*x + e  *arbconst(9) - 2*x

                                       1
  - 2*int(-----------------------------------------------------------,x)
                                                 2
                                                x
           sqrt(pi)*arbconst(9)*erf(i*x)*i*x + e  *arbconst(9) - 2*x

         2*y - 3
  - log(---------)=0}
            2


% (9) Integrating factors
% (FJW: Currently very slow, and fails anyway!)
% odesolve(sqrt(x)*df(y,x,2) + 2x*df(y,x) + 3y = 0, y, x);

% (10) Radical solution (FJW: omitted for now)

% (11) Undetermined coefficients
odesolve(df(y,x,2) - 2/x^2*y = 7x^4 + 3*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'.

                 2                 -1    1   6    1   5
{y=arbconst(13)*x  + arbconst(12)*x   + ---*x  + ---*x }
                                         4        6


% (12) Variation of parameters
odesolve(df(y,x,2) + y = csc(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'.
The Wronskian is 1

{y=arbconst(15)*sin(x) + arbconst(14)*cos(x) - cos(x)*x + log(sin(x))*sin(x)}


% (13) Linear constant coefficients
<< factor exp(x);  write
odesolve(df(y,x,7) - 14df(y,x,6) + 80df(y,x,5) - 242df(y,x,4)
   + 419df(y,x,3) - 416df(y,x,2) + 220df(y,x) - 48y = 0, y, x);
remfac exp(x) >>;

This is a linear ODE of order 7.
It has constant coefficients.

    4*x                 3*x                 2*x
{y=e   *arbconst(17) + e   *arbconst(16) + e   *(arbconst(19) + arbconst(18)*x)

     x                                                2
  + e *(arbconst(22) + arbconst(21)*x + arbconst(20)*x )}


% (14) Euler
odesolve(df(y,x,4) - 4/x^2*df(y,x,2) + 8/x^3*df(y,x) - 8/x^4*y = 0, y, x);

This is a linear ODE of order 4.
It has non-constant coefficients.
It is of the homogeneous (Euler) type and is reducible to a simpler ODE ...
It has constant coefficients.

                 4                 2                                  -1
{y=arbconst(26)*x  + arbconst(25)*x  + arbconst(24)*x + arbconst(23)*x  }


% (15) Exact n-th order
odesolve((1+x+x^2)*df(y,x,3) + (3+6x)*df(y,x,2) + 6df(y,x) = 6x, y, x);

This is a linear ODE of order 3.
It has non-constant coefficients.

Performing trivial order reduction to give the order 2

                                                         2
 linear ODE with coefficients (low -- high): {6,6*x + 3,x  + x + 1}

But ODESolve cannot solve the reduced ODE! 
It is exact, and the following linear ODE of order 2 is a first integral:

           2
df(y,x,2)*x  + df(y,x,2)*x + df(y,x,2) + 4*df(y,x)*x + 2*df(y,x) + 2*y

          2
=g34 + 3*x
It is exact, and the following linear ODE of order 1 is a first integral:

         2                                                  3
df(y,x)*x  + df(y,x)*x + df(y,x) + 2*x*y + y=g34*x + g35 + x
It is solved by the integrating factor method.

                                     1                2    1   4
    arbconst(29) + arbconst(28)*x + ---*arbconst(27)*x  + ---*x
                                     2                     4
{y=--------------------------------------------------------------}
                              2
                             x  + x + 1



% 1.2 Nonlinear equations
% =======================

% (16) Integrating factors 1
odesolve(df(y,x) = y/(y*log y + x), y, x);

This is a nonlinear ODE of order 1.
Interchanging dependent and independent variables ...

 - df(x,y)*y + log(y)*y + x

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

                     1        2
{x=arbconst(30)*y + ---*log(y) *y}
                     2


% (17) Integrating factors 2
odesolve(2y*df(y,x)^2 - 2x*df(y,x) - y = 0, y, x);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

               5                  3                          2
4*df(x,G39)*G39  - 8*df(x,G39)*G39  + 3*df(x,G39)*G39 + 4*G39 *x + 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                  1/3                 2    1/3
                               2*2   *arbconst(31)*G39  - 2   *arbconst(31)
The subsidiary solution is {x=----------------------------------------------}
                                        2/3       4         2     1/3
                                     G39   *(4*G39  - 12*G39  + 9)

 and the main ODE can be solved parametrically in terms of the derivative.


                    4                 2      - 1/3             - 2/3  1/3
{{y=2*(4*arbparam(1)  - 12*arbparam(1)  + 9)      *arbparam(1)      *2

  *arbconst(31)*arbparam(1),

                    4                 2      - 1/3             - 2/3  1/3
  x=2*(4*arbparam(1)  - 12*arbparam(1)  + 9)      *arbparam(1)      *2

                           2                 4                 2      - 1/3
  *arbconst(31)*arbparam(1)  - (4*arbparam(1)  - 12*arbparam(1)  + 9)

               - 2/3  1/3
  *arbparam(1)      *2   *arbconst(31),

  arbparam(1)}}

% This parametric solution is correct, cf. Zwillinger (1989) p.168 (41.10)
% (except that first edition is missing the constant C)!

% (18) Bernoulli 1
odesolve(df(y,x) + y = y^3*sin x, y, x, explicit);

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

{y

      2*x                                     - 1/2
 =(5*e   *arbconst(32) + 2*cos(x) + 4*sin(x))      *sqrt(5)*plus_or_minus(tag_1)

 }

expand_plus_or_minus ws;


       2*x                                     - 1/2
{y=(5*e   *arbconst(32) + 2*cos(x) + 4*sin(x))      *sqrt(5),

          2*x                                     - 1/2
 y= - (5*e   *arbconst(32) + 2*cos(x) + 4*sin(x))      *sqrt(5)}


% (19) Bernoulli 2
depend {P, Q}, x;


begin scalar soln, !*exp, !*allfac;  % for a neat solution
   on allfac;
   soln := odesolve(df(y,x) + P*y = Q*y^n, y, x);
   off allfac;  return soln
end;

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

                                                  int(p,x)
   - n       int(p,x)*n - int(p,x)               e        *q
{y    *y= - e                     *((n - 1)*int(-------------,x) - arbconst(33))
                                                  int(p,x)*n
                                                 e

 }

odesolve(df(y,x) + P*y = Q*y^(2/3), y, x);

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

  1/3   - 1/3*int(p,x)                 1    - 1/3*int(p,x)      int(p,x)/3
{y   =e               *arbconst(34) + ---*e               *int(e          *q,x)}
                                       3


% (20) Clairaut 1
odesolve((x^2-1)*df(y,x)^2 - 2x*y*df(y,x) + y^2 - 1 = 0, y, x, explicit);

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

Solution before trying to solve for dependent variable is 

            2  2               2                         2
arbconst(35) *x  - arbconst(35)  - 2*arbconst(35)*x*y + y  - 1=0


                                                              2    2
Solution before trying to solve for dependent variable is  - x  - y  + 1=0


                                     2
{y=arbconst(35)*x + sqrt(arbconst(35)  + 1),

                                     2
 y=arbconst(35)*x - sqrt(arbconst(35)  + 1),

            2
 y=sqrt( - x  + 1),

               2
 y= - sqrt( - x  + 1)}


% (21) Clairaut 2
operator f, g;


odesolve(f(x*df(y,x)-y) = g(df(y,x)), y, x);

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

{f(arbconst(36)*x - y) - g(arbconst(36))=0}


% (22) Equations of the form  y' = f(x,y)
odesolve(df(y,x) = (3x^2-y^2-7)/(exp(y)+2x*y+1), y, x);

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

                 y    3      2
{arbconst(37) + e  - x  + x*y  + 7*x + y=0}


% (23) Homogeneous
odesolve(df(y,x) = (2x^3*y-y^4)/(x^4-2x*y^3), 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'.

                     3    3
{arbconst(38)*x*y + x  + y =0}


% (24) Factoring the equation
odesolve(df(y,x)*(df(y,x)+y) = x*(x+y), y, x);

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

df(y,x) + x + y=0

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

df(y,x) - x=0

This is a linear ODE of order 1.
It is solved by quadrature.

     - x                                        1   2
{y=e    *arbconst(39) - x + 1,y=arbconst(40) + ---*x }
                                                2


% (25) Interchange variables
% (NB: Soln in Zwillinger (1989) wrong, as is last eqn in Table 68!)
odesolve(df(y,x) = x/(x^2*y^2+y^5), y, x);

This is a nonlinear ODE of order 1.
Interchanging dependent and independent variables ...

                2  2    5
 - df(x,y)*x + x *y  + y

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

          3
  2  2/3*y                  3    3
{x =e      *arbconst(41) - y  - ---}
                                 2


% (26) Lagrange 1
odesolve(y = 2x*df(y,x) - a*df(y,x)^3, y, x);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

                        2
 - df(x,G58)*G58 + 3*G58 *a - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                                     4
                               4*arbconst(42) + 3*G58 *a
The subsidiary solution is {x=---------------------------}
                                             2
                                        4*G58

 and the main ODE can be solved parametrically in terms of the derivative.


                              -1    1             3
{{y=2*arbconst(42)*arbparam(2)   + ---*arbparam(2) *a,
                                    2

                            -2    3             2
  x=arbconst(42)*arbparam(2)   + ---*arbparam(2) *a,
                                  4

  arbparam(2)}}

odesolve(y = 2x*df(y,x) - a*df(y,x)^3, y, x, implicit);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

                        2
 - df(x,G59)*G59 + 3*G59 *a - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                                     4
                               4*arbconst(43) + 3*G59 *a
The subsidiary solution is {x=---------------------------}
                                             2
                                        4*G59

 and the main ODE can be solved parametrically in terms of the derivative.


                3  2                   2    2                         2
{64*arbconst(43) *a  + 128*arbconst(43) *a*x  - 144*arbconst(43)*a*x*y

                     4         4       3  2
  + 64*arbconst(43)*x  + 27*a*y  - 16*x *y =0}

% root_of quartic is VERY slow if explicit option used!

% (27) Lagrange 2
odesolve(y = 2x*df(y,x) - df(y,x)^2, y, x);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

 - df(x,G60)*G60 + 2*G60 - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                                     3
                               3*arbconst(44) + 2*G60
The subsidiary solution is {x=-------------------------}
                                            2
                                       3*G60

 and the main ODE can be solved parametrically in terms of the derivative.


                              -1    1             2
{{y=2*arbconst(44)*arbparam(3)   + ---*arbparam(3) ,
                                    3

                            -2    2
  x=arbconst(44)*arbparam(3)   + ---*arbparam(3),
                                  3

  arbparam(3)}}

odesolve(y = 2x*df(y,x) - df(y,x)^2, y, x, implicit);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

 - df(x,G61)*G61 + 2*G61 - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                                     3
                               3*arbconst(45) + 2*G61
The subsidiary solution is {x=-------------------------}
                                            2
                                       3*G61

 and the main ODE can be solved parametrically in terms of the derivative.


                  2                    3                            2  2      3
{ - 9*arbconst(45)  - 12*arbconst(45)*x  + 18*arbconst(45)*x*y + 3*x *y  - 4*y

 =0}


% (28) Riccati 1
odesolve(df(y,x) = exp(x)*y^2 - y + exp(-x), y, x);

This is a nonlinear ODE of order 1.

It is of Riccati type and transforms into the linear second-order ODE: 

df(y,x,2) + y=0

It has constant coefficients.

      - x                         - x
    e    *arbconst(46)*sin(x) - e    *cos(x)
{y=------------------------------------------}
          arbconst(46)*cos(x) + sin(x)


% (29) Riccati 2
factor x;


odesolve(df(y,x) = y^2 - x*y + 1, y, x);

This is a nonlinear ODE of order 1.

It is of Riccati type and transforms into the linear second-order ODE: 

df(y,x,2) + df(y,x)*x + y=0

It has non-constant coefficients.
It is exact, and the following linear ODE of order 1 is a first integral:

df(y,x) + x*y=g66
It is solved by the integrating factor method.

                               2
                          1/2*x
                       2*e      *arbconst(47)
{y=x + ------------------------------------------------------}
                                            - 1/2
        sqrt(pi)*sqrt(2)*arbconst(47)*erf(2      *x*i)*i - 2

remfac x;



% (30) Separable
odesolve(df(y,x) = (9x^8+1)/(y^2+1), y, x);

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

                     9          3
{3*arbconst(48) - 3*x  - 3*x + y  + 3*y=0}


% (31) Solvable for x
odesolve(y = 2x*df(y,x) + y*df(y,x)^2, y, x);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

             5                        2
df(x,G68)*G68  - df(x,G68)*G68 - 2*G68 *x - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                               2
                               arbconst(49)*G68  - arbconst(49)
The subsidiary solution is {x=----------------------------------}
                                                2
                                             G68

 and the main ODE can be solved parametrically in terms of the derivative.


                                 -1
{{y= - 2*arbconst(49)*arbparam(4)  ,

                               -2
  x= - arbconst(49)*arbparam(4)   + arbconst(49),

  arbparam(4)}}

odesolve(y = 2x*df(y,x) + y*df(y,x)^2, y, x, implicit);

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

             5                        2
df(x,G69)*G69  - df(x,G69)*G69 - 2*G69 *x - 2*x=0

This is a first-order linear ODE solved by the integrating factor method.

                                               2
                               arbconst(50)*G69  - arbconst(50)
The subsidiary solution is {x=----------------------------------}
                                                2
                                             G69

 and the main ODE can be solved parametrically in terms of the derivative.


                  2                       2
{ - 4*arbconst(50)  + 4*arbconst(50)*x + y =0}


% (32) Solvable for y
begin scalar !*allfac;  !*allfac := t;  return
   odesolve(x = y*df(y,x) - x*df(y,x)^2, y, x)
end;

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

                      2
 - df(x,G70)*G70 - G70 *x + x=0

This is a first-order linear ODE solved by the integrating factor method.

                               arbconst(51)*G70
The subsidiary solution is {x=------------------}
                                       2
                                    G70 /2
                                   e

 and the main ODE can be solved parametrically in terms of the derivative.


                       2
      - 1/2*arbparam(5)                           2
{{y=e                   *arbconst(51)*(arbparam(5)  + 1),

                       2
      - 1/2*arbparam(5)
  x=e                   *arbconst(51)*arbparam(5),

  arbparam(5)}}


% (33) Autonomous 1
odesolve(df(y,x,2)-df(y,x) = 2y*df(y,x), y, x, explicit);

This is a nonlinear ODE of order 2.

This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: df(G71,y)*G71 - 2*G71*y - G71=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

df(G71,y) - 2*y - 1=0

This is a linear ODE of order 1.
It is solved by quadrature.
Restoring order to give these first-order ODEs ...

                             2
 - arbconst(52) + df(y,x) - y  - y=0

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

df(y,x)=0

This is a linear ODE of order 1.
It is solved by quadrature.

Simplifying the arbconst expressions in 4*arbconst(53)*arbconst(52)

 - arbconst(53) - 4*arbconst(52)*x

                                            2*y + 1
 + 2*sqrt(4*arbconst(52) - 1)*atan(--------------------------) + x
                                    sqrt(4*arbconst(52) - 1)

 by the rewrites ...


sqrt(4*arbconst(52) - 1) => arbconst(52)


Solution before trying to solve for dependent variable is 

                         2               2
arbconst(53)*arbconst(52)  - arbconst(52) *x

                          2*y + 1
 + 2*arbconst(52)*atan(--------------)=0
                        arbconst(52)


       1                    1                               1
{y= - ---*arbconst(52)*tan(---*arbconst(53)*arbconst(52) - ---*arbconst(52)*x)
       2                    2                               2

     1
  - ---,
     2

 y=arbconst(54)}


% (34) Autonomous 2  (FJW: Slow without either algint or noint option.)
odesolve(df(y,x,2)/y - df(y,x)^2/y^2 - 1 + 1/y^3 = 0, y, x, algint);

This is a nonlinear ODE of order 2.

This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

               2      2      3
df(G75,y)*G75*y  - G75 *y - y  + 1=0

This is a nonlinear ODE of order 1.
It is of Bernoulli type.
Restoring order to give these first-order ODEs ...

sqrt(y)*sqrt(3)*df(y,x)

                        3             3
 - sqrt(3*arbconst(56)*y  + 6*log(y)*y  + 2)*plus_or_minus(tag_4)=0

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

{arbconst(57)*plus_or_minus(tag_4)

                                  sqrt(y)
  + sqrt(3)*int(-------------------------------------------,y)
                                      3             3
                 sqrt(3*arbconst(56)*y  + 6*log(y)*y  + 2)

  - plus_or_minus(tag_4)*x=0}


% (35) Differentiation method
odesolve(2y*df(y,x,2) - df(y,x)^2 = 1/3(df(y,x) - x*df(y,x,2))^2, y, x, explicit);

This is a nonlinear ODE of order 2.

                                                                            G84
This ODE is equidimensional in the independent variable x -- applying x => e

                                                2
 to transform to the simpler ODE:  - df(y,G84,2)  + 4*df(y,G84,2)*df(y,G84)

                                2
 + 6*df(y,G84,2)*y - 7*df(y,G84)  - 6*df(y,G84)*y=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

            2    2                  2                            2
 - df(G85,y) *G85  + 4*df(G85,y)*G85  + 6*df(G85,y)*G85*y - 7*G85  - 6*G85*y=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

            2
 - df(G85,y) *G85 + 4*df(G85,y)*G85 + 6*df(G85,y)*y - 7*G85 - 6*y=0

This is a nonlinear ODE of order 1.

It is of Lagrange type and reduces to this subsidiary ODE for x(y'): 

             5                  4                   3                   2
df(y,G86)*G86  - 8*df(y,G86)*G86  + 24*df(y,G86)*G86  - 26*df(y,G86)*G86

                                          2
 - 17*df(y,G86)*G86 + 42*df(y,G86) + 6*G86 *y - 12*G86*y - 18*y=0

This is a first-order linear ODE solved by the integrating factor method.

The subsidiary solution is {y

                     2
     arbconst(58)*G86  - 4*arbconst(58)*G86 + 7*arbconst(58)
   =---------------------------------------------------------}
                           2
                        G86  - 4*G86 + 4

 and the main ODE can be solved parametrically in terms of the derivative.

Restoring order to give these first-order ODEs ...

               2                                                           2
16*arbconst(58)  + 4*arbconst(58)*df(y,G84) - 20*arbconst(58)*y + df(y,G84)

                      2
 - 4*df(y,G84)*y + 4*y =0

This is a nonlinear ODE of order 1.
It can be (partially) solved algebraically for the single-order derivative 
and each `root ODE' will be solved separately ...

2*arbconst(58) + df(y,G84)

 - 2*sqrt( - arbconst(58) + y)*sqrt(arbconst(58))*sqrt(3) - 2*y=0

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

2*arbconst(58) + df(y,G84)

 + 2*sqrt( - arbconst(58) + y)*sqrt(arbconst(58))*sqrt(3) - 2*y=0

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

df(y,G84)=0

This is a linear ODE of order 1.
It is solved by quadrature.

                                            G84
Simplifying the arbconst expressions in  - e   *arbconst(59)

 - sqrt(arbconst(58))*sqrt(3) - sqrt( - arbconst(58) + y) by the rewrites ...


sqrt(arbconst(58)) => arbconst(58)


                                            G84
Simplifying the arbconst expressions in  - e   *arbconst(60)

 + sqrt(arbconst(58))*sqrt(3) - sqrt( - arbconst(58) + y) by the rewrites ...


sqrt(arbconst(58)) => arbconst(58)


Solution before trying to solve for dependent variable is 

                                                               2
 - arbconst(59)*x - sqrt(3)*arbconst(58) - sqrt( - arbconst(58)  + y)=0


Solution before trying to solve for dependent variable is 

                                                               2
 - arbconst(60)*x + sqrt(3)*arbconst(58) - sqrt( - arbconst(58)  + y)=0


               2  2                                                         2
{y=arbconst(59) *x  + 2*sqrt(3)*arbconst(59)*arbconst(58)*x + 4*arbconst(58) ,

               2  2                                                         2
 y=arbconst(60) *x  - 2*sqrt(3)*arbconst(60)*arbconst(58)*x + 4*arbconst(58) ,

 y=arbconst(61)}


% (36) Equidimensional in x
odesolve(x*df(y,x,2) = 2y*df(y,x), y, x, explicit);

This is a nonlinear ODE of order 2.

                                                                            G95
This ODE is equidimensional in the independent variable x -- applying x => e

 to transform to the simpler ODE: df(y,G95,2) - 2*df(y,G95)*y - df(y,G95)=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: df(G96,y)*G96 - 2*G96*y - G96=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

df(G96,y) - 2*y - 1=0

This is a linear ODE of order 1.
It is solved by quadrature.
Restoring order to give these first-order ODEs ...

                               2
 - arbconst(62) + df(y,G95) - y  - y=0

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

df(y,G95)=0

This is a linear ODE of order 1.
It is solved by quadrature.

Simplifying the arbconst expressions in 4*arbconst(63)*arbconst(62)

 - arbconst(63) - 4*arbconst(62)*G95

                                            2*y + 1
 + 2*sqrt(4*arbconst(62) - 1)*atan(--------------------------) + G95
                                    sqrt(4*arbconst(62) - 1)

 by the rewrites ...


sqrt(4*arbconst(62) - 1) => arbconst(62)


Solution before trying to solve for dependent variable is 

                         2               2
arbconst(63)*arbconst(62)  - arbconst(62) *log(x)

                          2*y + 1
 + 2*arbconst(62)*atan(--------------)=0
                        arbconst(62)


       1
{y= - ---*arbconst(62)
       2

       1                               1                          1
 *tan(---*arbconst(63)*arbconst(62) - ---*arbconst(62)*log(x)) - ---,
       2                               2                          2

 y=arbconst(64)}


% (37) Equidimensional in y
odesolve((1-x)*(y*df(y,x,2)-df(y,x)^2) + x^2*y^2 = 0, y, x);

This is a nonlinear ODE of order 2.

                                                                          G102
This ODE is equidimensional in the dependent variable y -- applying y => e

                                                                      2
 to transform to the simpler ODE:  - df(G102,x,2)*x + df(G102,x,2) + x =0

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

                                          3        2
     arbconst(66) + arbconst(65)*x + 1/6*x  + 1/2*x  - x        x
    e                                                   *(x - 1)
{y=---------------------------------------------------------------}
                                x - 1


% (38) Exact second order
odesolve(x*y*df(y,x,2) + x*df(y,x)^2 + y*df(y,x) = 0, y, x, explicit);

This is a nonlinear ODE of order 2.

                                                                            G106
This ODE is equidimensional in the independent variable x -- applying x => e

                                                             2
 to transform to the simpler ODE: df(y,G106,2)*y + df(y,G106) =0


This ODE is autonomous -- transforming dependent variable 

                                                                         2
to derivative to give this ODE of order 1 lower: df(G107,y)*G107*y + G107 =0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

df(G107,y)*y + G107=0

This is a linear ODE of order 1.
It is solved by the integrating factor method.
Restoring order to give these first-order ODEs ...

 - arbconst(67) + df(y,G106)*y=0

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

df(y,G106)=0

This is a linear ODE of order 1.
It is solved by quadrature.

Solution before trying to solve for dependent variable is 

                                                       2
2*arbconst(68)*arbconst(67) - 2*arbconst(67)*log(x) + y =0


{y=sqrt( - arbconst(68) + log(x))*sqrt(arbconst(67))*sqrt(2),

 y= - sqrt( - arbconst(68) + log(x))*sqrt(arbconst(67))*sqrt(2),

 y=arbconst(69)}


% (39) Factoring differential operator
odesolve(df(y,x,2)^2 - 2df(y,x)*df(y,x,2) + 2y*df(y,x) - y^2 = 0, y, x);

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

df(y,x,2) - 2*df(y,x) + y=0

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

df(y,x,2) - y=0

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

    x                 x
{y=e *arbconst(71) + e *arbconst(70)*x,

    x                  - x
 y=e *arbconst(73) + e    *arbconst(72)}


% (40) Scale invariant (fails with algint option)
odesolve(x^2*df(y,x,2) + 3x*df(y,x) = 1/(y^3*x^4), y, x);

This is a nonlinear ODE of order 2.

                                              -1
This ODE is scale invariant -- applying y => x  *G113

 to transform to the simpler ODE: 

                 3  2                  3         4
df(G113,x,2)*G113 *x  + df(G113,x)*G113 *x - G113  - 1=0


                                                                            G115
This ODE is equidimensional in the independent variable x -- applying x => e

                                                      3       4
 to transform to the simpler ODE: df(G113,G115,2)*G113  - G113  - 1=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

                  3            4
df(G116,G113)*G113 *G116 - G113  - 1=0

This is a nonlinear ODE of order 1.
It is separable.
Restoring order to give these first-order ODEs ...

                   2                2     2       4
4*arbconst(74)*G113  + df(G113,G115) *G113  - G113  + 1=0

This is a nonlinear ODE of order 1.
It can be (partially) solved algebraically for the single-order derivative 
and each `root ODE' will be solved separately ...

df(G113,G115)*G113

                              2       4
 - sqrt( - 4*arbconst(74)*G113  + G113  - 1)*plus_or_minus(tag_6)=0

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

{2*arbconst(75)*plus_or_minus(tag_6)

                                                     2  2    4  4         2  2
          - 2*arbconst(74) + sqrt( - 4*arbconst(74)*x *y  + x *y  - 1) + x *y
  + log(-----------------------------------------------------------------------)
                                                  2
                               sqrt(4*arbconst(74)  + 1)

  - 2*log(x)*plus_or_minus(tag_6)=0}


% Revised scale-invariant example (hangs with algint option):
ode := x^2*df(y,x,2) + 3x*df(y,x) + 2*y = 1/(y^3*x^4);


                  2                      -4  -3
ode := df(y,x,2)*x  + 3*df(y,x)*x + 2*y=x  *y

% Choose full (explicit and expanded) solution:
odesolve(ode, y, x, full);

This is a nonlinear ODE of order 2.

                                              -1
This ODE is scale invariant -- applying y => x  *G122

 to transform to the simpler ODE: 

                 3  2                  3         4
df(G122,x,2)*G122 *x  + df(G122,x)*G122 *x + G122  - 1=0


                                                                            G124
This ODE is equidimensional in the independent variable x -- applying x => e

                                                      3       4
 to transform to the simpler ODE: df(G122,G124,2)*G122  + G122  - 1=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

                  3            4
df(G125,G122)*G122 *G125 + G122  - 1=0

This is a nonlinear ODE of order 1.
It is separable.
Restoring order to give these first-order ODEs ...

                    2                  2     2         4
15*arbconst(76)*G122  - 4*df(G122,G124) *G122  - 4*G122  - 4=0

This is a nonlinear ODE of order 1.
It can be (partially) solved algebraically for the single-order derivative 
and each `root ODE' will be solved separately ...

2*df(G122,G124)*G122

                            2         4
 - sqrt(15*arbconst(76)*G122  - 4*G122  - 4)*plus_or_minus(tag_7)=0

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

Solution before trying to solve for dependent variable is 

                                                                  2  2
                                             15*arbconst(76) - 8*x *y
2*arbconst(77)*plus_or_minus(tag_7) - asin(------------------------------)
                                                                 2
                                            sqrt(225*arbconst(76)  - 64)

 - 2*log(x)*plus_or_minus(tag_7)=0


       1
{y= - ---*sqrt(15*arbconst(76)
       2

                                       2
                - sqrt(225*arbconst(76)  - 64)*sin(2*arbconst(77) - 2*log(x)))

    - 1/2  -1
 *2      *x  ,

       1
 y= - ---*sqrt(15*arbconst(76)
       2

                                       2
                + sqrt(225*arbconst(76)  - 64)*sin(2*arbconst(77) - 2*log(x)))

    - 1/2  -1
 *2      *x  ,

    1
 y=---*sqrt(15*arbconst(76)
    2

                                    2
             - sqrt(225*arbconst(76)  - 64)*sin(2*arbconst(77) - 2*log(x)))

    - 1/2  -1
 *2      *x  ,

    1
 y=---*sqrt(15*arbconst(76)
    2

                                    2
             + sqrt(225*arbconst(76)  - 64)*sin(2*arbconst(77) - 2*log(x)))

    - 1/2  -1
 *2      *x  }
              % or "explicit, expand"
% Check it -- each solution should simplify to 0:
foreach soln in ws collect
   trigsimp sub(soln, num(lhs ode - rhs ode));


{0,0,0,0}


% (41) Autonomous, 3rd order
odesolve((df(y,x)^2+1)*df(y,x,3) - 3df(y,x)*df(y,x,2)^2 = 0, y, x);

This is a nonlinear ODE of order 3.

Performing trivial order reduction to give the order 2 nonlinear ODE: 

           2                        2
df(y,x,2)*y  + df(y,x,2) - 3*df(y,x) *y=0


This ODE is autonomous -- transforming dependent variable 

to derivative to give this ODE of order 1 lower: 

                 2                           2
df(G131,y)*G131*y  + df(G131,y)*G131 - 3*G131 *y=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

            2
df(G131,y)*y  + df(G131,y) - 3*G131*y=0

This is a linear ODE of order 1.
It is solved by the integrating factor method.
Restoring order to give these first-order ODEs ...

         2                    2         2
 - sqrt(y  + 1)*arbconst(78)*y  - sqrt(y  + 1)*arbconst(78) + df(y,x)=0

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

df(y,x)=0

This is a linear ODE of order 1.
It is solved by quadrature.

                                                             2
Solution of order-reduced ODE is {arbconst(79)*arbconst(78)*y

                                                  2
    + arbconst(79)*arbconst(78) - arbconst(78)*x*y  - arbconst(78)*x

            2           2
    + sqrt(y  + 1)*y + y  + 1=0,

   y=arbconst(80)}


                                                                            2
Restoring order, y => df(y,x,1), to give: {arbconst(79)*arbconst(78)*df(y,x)

                                                      2
    + arbconst(79)*arbconst(78) - arbconst(78)*df(y,x) *x - arbconst(78)*x

             2               2
    + df(y,x)  + sqrt(df(y,x)  + 1)*df(y,x) + 1=0,

   df(y,x)=arbconst(80)} and re-solving ...


Simplifying the arbconst expressions in (arbconst(81)*arbconst(78) + sqrt(

                  2
      arbconst(79) *arbconst(78) - 2*arbconst(79)*arbconst(78)*x

                                        2
       + 2*arbconst(79) + arbconst(78)*x  - 2*x)*sqrt(arbconst(78))*i)/arbconst(

   78) by the rewrites ...


sqrt(arbconst(78)) => arbconst(78)


                                   2             2
{y=arbconst(81) + sqrt(arbconst(79) *arbconst(78)

                                  2                                  2  2
     - 2*arbconst(79)*arbconst(78) *x + 2*arbconst(79) + arbconst(78) *x  - 2*x)

              -1
 *arbconst(78)  *i,

 y=arbconst(82) + i*x,

 y=arbconst(83) - i*x,

 y=arbconst(84) + arbconst(80)*x}


% (42) Autonomous, 4th order
odesolve(3*df(y,x,2)*df(y,x,4) - 5df(y,x,3)^2 = 0, y, x);

This is a nonlinear ODE of order 4.

Performing trivial order reduction to give the order 2 nonlinear ODE: 

                         2
3*df(y,x,2)*y - 5*df(y,x) =0


This ODE is autonomous -- transforming dependent variable 

                                                                             2
to derivative to give this ODE of order 1 lower: 3*df(G136,y)*G136*y - 5*G136 =0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

3*df(G136,y)*y - 5*G136=0

This is a linear ODE of order 1.
It is solved by the integrating factor method.
Restoring order to give these first-order ODEs ...

    2/3
 - y   *arbconst(85)*y + df(y,x)=0

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

df(y,x)=0

This is a linear ODE of order 1.
It is solved by quadrature.

Solution of order-reduced ODE is {

      2/3                                2/3
   2*y   *arbconst(86)*arbconst(85) - 2*y   *arbconst(85)*x - 3=0,

   y=arbconst(87)}


Restoring order, y => df(y,x,2), to give: {

              2/3                                        2/3
   2*df(y,x,2)   *arbconst(86)*arbconst(85) - 2*df(y,x,2)   *arbconst(85)*x - 3

   =0,

   df(y,x,2)=arbconst(87)} and re-solving ...


                                                                  2
Simplifying the arbconst expressions in (arbconst(89)*arbconst(85) *x

                               2
    + arbconst(88)*arbconst(85)

                                                                       2
    - 3*sqrt(arbconst(86) - x)*sqrt(arbconst(85))*sqrt(6))/arbconst(85)

 by the rewrites ...


sqrt(arbconst(85)) => arbconst(85)


                                                                  2
Simplifying the arbconst expressions in (arbconst(91)*arbconst(85) *x

                               2
    + arbconst(90)*arbconst(85)

                                                                       2
    + 3*sqrt(arbconst(86) - x)*sqrt(arbconst(85))*sqrt(6))/arbconst(85)

 by the rewrites ...


sqrt(arbconst(85)) => arbconst(85)


{y=arbconst(89)*x + arbconst(88)

                                                 -3
  - 3*sqrt(arbconst(86) - x)*sqrt(6)*arbconst(85)  ,

 y=arbconst(91)*x + arbconst(90)

                                                 -3
  + 3*sqrt(arbconst(86) - x)*sqrt(6)*arbconst(85)  ,

                                    1                2
 y=arbconst(93)*x + arbconst(92) + ---*arbconst(87)*x }
                                    2



% 1.3 Special equations
% =====================

% (43) Delay
operator y;


odesolve(df(y(x),x) + a*y(x-1) = 0, y(x), x);


***** Arguments of y differ -- solving delay equations is not implemented.


% (44) Functions with several parameters
odesolve(df(y(x,a),x) = a*y(x,a), y(x,a), x);

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

         a*x
{y(x,a)=e   *arbconst(94)}



% 2  Single equations with initial conditions
% ===========================================

% (45) Exact 4th order
odesolve(df(y,x,4) = sin x, y, x,
   {x=0, y=0, df(y,x)=0, df(y,x,2)=0, df(y,x,3)=0});

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

General solution is {y

                                                  2                 3
   =arbconst(98) + arbconst(97)*x + arbconst(96)*x  + arbconst(95)*x  + sin(x)}


Applying conditions {{x=0,y=0,df(y,x)=0,df(y,x,2)=0,df(y,x,3)=0}}


             1   3
{y=sin(x) + ---*x  - x}
             6


% (46) Linear polynomial coefficients -- Bessel J0
odesolve(x*df(y,x,2) + df(y,x) + 2x*y = 0, y, x,
   {x=0, y=1, df(y,x)=0});

This is a linear ODE of order 2.
It has non-constant coefficients.
The reduced ODE can be solved in terms of special functions.

General solution is {y

   =arbconst(100)*bessely(0,sqrt(2)*x) + arbconst(99)*besselj(0,sqrt(2)*x)}


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


{y=besselj(0,sqrt(2)*x)}


% (47) Second-degree separable
soln :=
   odesolve(x*df(y,x)^2 - y^2 + 1 = 0, y=1, x=0, explicit);

This is a nonlinear ODE of order 1.
It can be (partially) solved algebraically for the single-order derivative 
and each `root ODE' will be solved separately ...

                        2
sqrt(x)*df(y,x) - sqrt(y  - 1)*plus_or_minus(tag_8)=0

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

General solution is {arbconst(101)*plus_or_minus(tag_8)

                                                 2
    - 2*sqrt(x)*plus_or_minus(tag_8) + log(sqrt(y  - 1) + y)=0}


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


Solution before trying to solve for dependent variable is 

                                              2
 - 2*sqrt(x)*plus_or_minus(tag_8) + log(sqrt(y  - 1) + y)=0


soln := 

    1   2*sqrt(x)*plus_or_minus(tag_8)    1    - 2*sqrt(x)*plus_or_minus(tag_8)
{y=---*e                               + ---*e                                 }
    2                                     2

% Alternatively ...
soln where e^~x => cosh x + sinh x;


{y=cosh(2*sqrt(x)*plus_or_minus(tag_8))}

% but this works ONLY with `on div, intstr; off allfac;'
% A better alternative is ...
trigsimp(soln, hyp, combine);


{y=cosh(2*sqrt(x)*plus_or_minus(tag_8))}

expand_plus_or_minus ws;


{y=cosh(2*sqrt(x))}


% (48) Autonomous
odesolve(df(y,x,2) + y*df(y,x)^3 = 0, y, x, {x=0, y=0, df(y,x)=2});

This is a nonlinear ODE of order 2.

This ODE is autonomous -- transforming dependent variable 

                                                                       3
to derivative to give this ODE of order 1 lower: df(G148,y)*G148 + G148 *y=0

This is a nonlinear ODE that factorizes algebraically 
and each distinct factor ODE will be solved separately ...

                 2
df(G148,y) + G148 *y=0

This is a nonlinear ODE of order 1.
It is separable.
Restoring order to give these first-order ODEs ...

                                   2
2*arbconst(102)*df(y,x) - df(y,x)*y  + 2=0

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

df(y,x)=0

This is a linear ODE of order 1.
It is solved by quadrature.

                                                                  3
General solution is {6*arbconst(103) - 6*arbconst(102)*y - 6*x + y =0,y

   =arbconst(104)}


Applying conditions {{x=0,y=0,df(y,x)=2}}


           3
{ - 6*x + y  + 3*y=0}

%% Only one explicit solution satisfies the conditions:
begin scalar !*trode, !*fullroots;  !*fullroots := t;  return
   odesolve(df(y,x,2) + y*df(y,x)^3 = 0, y, x,
      {x=0, y=0, df(y,x)=2}, explicit);
end;


            2            1/3            2             - 1/3
{y=(sqrt(9*x  + 1) + 3*x)    - (sqrt(9*x  + 1) + 3*x)      }



% 3  Systems of equations
% =======================

% (49) Integrable combinations
depend {x, y, z}, t;


odesolve({df(x,t) = -3y*z, df(y,t) = 3x*z, df(z,t) = -x*y}, {x,y,z}, t);


odesolve-system({df(x,t) + 3*y*z,

                 df(y,t) - 3*x*z,

                 df(z,t) + x*y},{x,y,z},t)


% (50) Matrix Riccati
depend {a, b}, t;


odesolve({df(x,t) = a*(y^2-x^2) + 2b*x*y + 2c*x,
   df(y,t) = b*(y^2-x^2) - 2a*x*y + 2c*y}, {x,y}, t);


                              2      2
odesolve-system({df(x,t) + a*x  - a*y  - 2*b*x*y - 2*c*x,

                                        2      2
                 df(y,t) + 2*a*x*y + b*x  - b*y  - 2*c*y},{x,y},t)


% (51) Triangular
odesolve({df(x,t) = x*(1 + cos(t)/(2+sin(t))), df(y,t) = x - y},
   {x,y}, t);


                   - cos(t)*x + df(x,t)*sin(t) + 2*df(x,t) - sin(t)*x - 2*x
odesolve-system({-----------------------------------------------------------,
                                         sin(t) + 2

                 df(y,t) - x + y},{x,y},t)


% (52) Vector
odesolve({df(x,t) = 9x + 2y, df(y,t) = x + 8y}, {x,y}, t);


odesolve-system({df(x,t) - 9*x - 2*y,df(y,t) - x - 8*y},{x,y},t)


% (53) Higher order
odesolve({df(x,t) - x + 2y = 0, df(x,t,2) - 2df(y,t) = 2t - cos(2t)},
   {x,y}, t);


odesolve-system({df(x,t) - x + 2*y,

                 cos(2*t) + df(x,t,2) - 2*df(y,t) - 2*t},{x,y},t)


% (54) Inhomogeneous system
equ := {df(x,t) = -1/(t*(t^2+1))*x + 1/(t^2*(t^2+1))*y + 1/t,
        df(y,t) = -t^2/(t^2+1)*x + (2t^2+1)/(t*(t^2+1))*y + 1};


                      -1      -1    -2
                 t - t  *x + t   + t  *y
equ := {df(x,t)=-------------------------,
                          2
                         t  + 1

                     2      2            -1
                  - t *x + t  + 2*t*y + t  *y + 1
        df(y,t)=----------------------------------}
                               2
                              t  + 1

odesolve(equ, {x,y}, t);


                           2                  -1      -1    -2
                  df(x,t)*t  + df(x,t) - t + t  *x - t   - t  *y
odesolve-system({------------------------------------------------,
                                       2
                                      t  + 1

                           2              2      2            -1
                  df(y,t)*t  + df(y,t) + t *x - t  - 2*t*y - t  *y - 1
                 ------------------------------------------------------},{x,y},t
                                          2
                                         t  + 1

)


end;



Time for test: 20727 ms, plus GC time: 1952 ms