arithmetic/src/algebra/root/RootBasic.ig


GENERIC INTERFACE RootBasic(R, P);
*Arithmetic for Modula-3, see doc for details

Abstract: Arithmetic on the roots of polynomials

E.g. from given polynomials x and y with the roots X={x1,...,xn} and Y={y1,...,ym} respectively, we can compute the polynomials that have the roots {xi+eta : xi in X and eta in Y} {xi*eta : xi in X and eta in Y} {xi^n : xi in X} etc. *


FROM Arithmetic IMPORT Error;

CONST Brand = R.Brand & "Root";

TYPE
  (* interpretation is: a[0] + a[1]*xi + a[2]* xi^2...a[n]*xi^n *)

  TBody = P.TBody;
  T = P.T;
*It's not possible to obtain a pointer to a constant array. We can not turn T from a reference type to an array type, because some routines have to return a result via a VAR parameter. CONST Zero = TBody{R.Zero,R.One}; One = TBody{R.MinusOne,R.One}; *

VAR
  (*CONST*)
  Zero: T;
  (*CONST*)
  One: T;

CONST New = P.New;
CONST Copy = P.Copy;

<* INLINE *>
PROCEDURE IsZero (x: T; ): BOOLEAN;
CONST Equal = P.Equal;

PROCEDURE Add (x, y: T; ): T;    (* x+y *)
PROCEDURE Sub (x, y: T; ): T;    (* x-y *)
PROCEDURE Neg (x: T; ): T;       (* -x *)

PROCEDURE Mul (x, y: T; ): T;    (* x*y *)
PROCEDURE Div (x, y: T; ): T
  RAISES {Error};                (* x/y if possible, will fail for floating
                                    point numbers often *)
PROCEDURE Rec (READONLY x: T; ): T RAISES {Error}; (* 1/x *)
PROCEDURE Mod (x, y: T; ): T RAISES {Error}; (* x mod y *)
PROCEDURE DivMod (x, y: T;  (* compute x/y *)VAR r: T; ): T
  RAISES {Error};                (* giving quotient with remainder r
                                    (always zero)*)

TYPE RootArray = ARRAY OF R.T;

PROCEDURE FromRoots (READONLY root: RootArray; ): T;
Create polynomial which has the roots as specified in array 'root'

PROCEDURE ElimMultRoots(x:T;):T; Eliminate multiple roots Can be used for simplifying polynomials after some basic operations.

PROCEDURE PowN (x: T; y: CARDINAL; ): T; (* x^y *)
PROCEDURE PowNSlow (x: T; y: CARDINAL; ): T; (* x^y *)
The number of roots doesn't change, thus PowN(x,2) is very different from Mul(x,x) because in the latter case every root is multiplied with _each_ other.

CONST RootOf = P.Compose;

TYPE PowerSumSeq = ARRAY OF R.T;

PROCEDURE ToPowerSumSeq (x: T; m: CARDINAL; ): REF PowerSumSeq;
For a given set of roots x1,...,xn (represented by a polynomial) calculate the sum x1^j+...+xn^j for j from 1 to m If m=n you receive enough values to reconstruct the polynomials later. Since for any of xi holds 0=a[0]+a[1]*xi+a[2]*xi^2+...+a[n]*xi^n, which means a[n]*xi^n=-(a[0]+a[1]*xi+a[2]*xi^2+...+a[n-1]*xi^(n-1)), one can continue the sequence of these power sums by a linear recurrence.

PROCEDURE FromPowerSumSeq (READONLY x: PowerSumSeq; ): T RAISES {Error};
Inverse to ToPowerSumSeq Err.indivisible is raised if x is not a sequence of power sums

PROCEDURE Deflate(x:T; (* divide this polynomial

c:R.T;      (* by (xi-c) *)
                  VAR rem:R.T;);(* leaving remainder -- possibly 0 *)
PROCEDURE Discriminant(x:T;        (* product of all possible differences of roots *)
                      ):T;
*)

END RootBasic.