aboutsummaryrefslogtreecommitdiff
path: root/src/algebra/oderf.spad.pamphlet
blob: 048606cef8fdbfd797f78be7ad44749b789578f9 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
379
380
381
382
383
384
385
386
387
388
389
390
391
392
393
394
395
396
397
398
399
400
401
402
403
404
405
406
407
408
409
410
411
412
413
414
415
416
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
440
441
442
443
444
445
446
447
448
449
450
451
452
453
454
455
456
457
458
459
460
461
462
463
464
465
466
467
468
469
470
471
472
473
474
475
476
477
478
479
480
481
482
483
484
485
486
487
488
489
490
491
492
493
494
495
496
497
498
499
500
501
502
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
525
526
527
528
529
530
531
532
533
534
535
536
537
538
539
540
541
542
543
544
545
546
547
548
549
550
551
552
553
554
555
556
557
558
559
560
561
562
563
564
565
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
830
831
832
833
834
835
836
837
838
839
840
841
842
843
844
845
846
847
848
849
850
851
852
853
854
855
856
857
858
859
860
861
862
863
864
865
866
867
868
869
870
871
872
873
874
875
876
877
878
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
\documentclass{article}
\usepackage{open-axiom}
\begin{document}
\title{\$SPAD/src/algebra oderf.spad}
\author{Manuel Bronstein}
\maketitle
\begin{abstract}
\end{abstract}
\eject
\tableofcontents
\eject
\section{package BALFACT BalancedFactorisation}
<<package BALFACT BalancedFactorisation>>=
)abbrev package BALFACT BalancedFactorisation
++ Author: Manuel Bronstein
++ Date Created: 1 March 1991
++ Date Last Updated: 11 October 1991
++ Description: This package provides balanced factorisations of polynomials.
BalancedFactorisation(R, UP): Exports == Implementation where
  R  : Join(GcdDomain, CharacteristicZero)
  UP : UnivariatePolynomialCategory R

  Exports ==> with
    balancedFactorisation: (UP, UP) -> Factored UP
      ++ balancedFactorisation(a, b) returns
      ++ a factorisation \spad{a = p1^e1 ... pm^em} such that each
      ++ \spad{pi} is balanced with respect to b.
    balancedFactorisation: (UP, List UP) -> Factored UP
      ++ balancedFactorisation(a, [b1,...,bn]) returns
      ++ a factorisation \spad{a = p1^e1 ... pm^em} such that each
      ++ pi is balanced with respect to \spad{[b1,...,bm]}.

  Implementation ==> add
    balSqfr : (UP, Integer, List UP) -> Factored UP
    balSqfr1: (UP, Integer,      UP) -> Factored UP

    balancedFactorisation(a:UP, b:UP) == balancedFactorisation(a, [b])

    balSqfr1(a, n, b) ==
      g := gcd(a, b)
      fa := sqfrFactor((a exquo g)::UP, n)
      ground? g => fa
      fa * balSqfr1(g, n, (b exquo (g ** order(b, g)))::UP)

    balSqfr(a, n, l) ==
      b := first l
      empty? rest l => balSqfr1(a, n, b)
      */[balSqfr1(f.factor, n, b) for f in factors balSqfr(a,n,rest l)]

    balancedFactorisation(a:UP, l:List UP) ==
      empty?(ll := select(#1 ~= 0, l)) =>
        error "balancedFactorisation: 2nd argument is empty or all 0"
      sa := squareFree a
      unit(sa) * */[balSqfr(f.factor,f.exponent,ll) for f in factors sa]

@
\section{package BOUNDZRO BoundIntegerRoots}
<<package BOUNDZRO BoundIntegerRoots>>=
)abbrev package BOUNDZRO BoundIntegerRoots
++ Author: Manuel Bronstein
++ Date Created: 11 March 1991
++ Date Last Updated: 18 November 1991
++ Description:
++   \spadtype{BoundIntegerRoots} provides functions to
++   find lower bounds on the integer roots of a polynomial.
BoundIntegerRoots(F, UP): Exports == Implementation where
  F  : Join(Field, RetractableTo Fraction Integer)
  UP : UnivariatePolynomialCategory F

  Z   ==> Integer
  Q   ==> Fraction Z
  K   ==> Kernel F
  UPQ ==> SparseUnivariatePolynomial Q
  ALGOP ==> "%alg"

  Exports ==> with
    integerBound: UP -> Z
      ++ integerBound(p) returns a lower bound on the negative integer
      ++ roots of p, and 0 if p has no negative integer roots.

  Implementation ==> add
    import RationalFactorize(UPQ)
    import UnivariatePolynomialCategoryFunctions2(F, UP, Q, UPQ)

    qbound : (UP, UPQ) -> Z
    zroot1 : UP -> Z
    qzroot1: UPQ -> Z
    negint : Q -> Z

-- returns 0 if p has no integer root < 0, its negative integer root otherwise
    qzroot1 p == negint(- leadingCoefficient(reductum p) / leadingCoefficient p)

-- returns 0 if p has no integer root < 0, its negative integer root otherwise
    zroot1 p ==
      z := - leadingCoefficient(reductum p) / leadingCoefficient p
      (r := retractIfCan(z)@Union(Q, "failed")) case Q => negint(r::Q)
      0

-- returns 0 if r is not a negative integer, r otherwise
    negint r ==
      ((u := retractIfCan(r)@Union(Z, "failed")) case Z) and negative?(u::Z) => u::Z
      0

    if F has ExpressionSpace then
      bringDown: F -> Q

-- the random substitution used by bringDown is NOT always a ring-homorphism
-- (because of potential algebraic kernels), but is ALWAYS a Z-linear map.
-- this guarantees that bringing down the coefficients of (x + n) q(x) for an
-- integer n yields a polynomial h(x) which is divisible by x + n
-- the only problem is that evaluating with random numbers can cause a
-- division by 0. We should really be able to trap this error later and
-- reevaluate with a new set of random numbers    MB 11/91
      bringDown f ==
        t := tower f
        retract eval(f, t, [random()$Q :: F for k in t])

      integerBound p ==
        one? degree p => zroot1 p
        q1 := map(bringDown, p)
        q2 := map(bringDown, p)
        qbound(p, gcd(q1, q2))

    else
      integerBound p ==
        one? degree p => zroot1 p
        qbound(p, map(retract(#1)@Q, p))

-- we can probably do better here (i.e. without factoring)
    qbound(p, q) ==
      bound:Z := 0
      for rec in factors factor q repeat
        if one?(degree(rec.factor)) and ((r := qzroot1(rec.factor)) < bound)
           and zero? p(r::Q::F) then bound := r
      bound

@
\section{package ODEPRIM PrimitiveRatDE}
<<package ODEPRIM PrimitiveRatDE>>=
)abbrev package ODEPRIM PrimitiveRatDE
++ Author: Manuel Bronstein
++ Date Created: 1 March 1991
++ Date Last Updated: 1 February 1994
++ Description:
++  \spad{PrimitiveRatDE} provides functions for in-field solutions of linear
++   ordinary differential equations, in the transcendental case.
++   The derivation to use is given by the parameter \spad{L}.
PrimitiveRatDE(F, UP, L, LQ): Exports == Implementation where
  F  : Join(Field, CharacteristicZero, RetractableTo Fraction Integer)
  UP : UnivariatePolynomialCategory F
  L  : LinearOrdinaryDifferentialOperatorCategory UP
  LQ : LinearOrdinaryDifferentialOperatorCategory Fraction UP

  N   ==> NonNegativeInteger
  Z   ==> Integer
  RF  ==> Fraction UP
  UP2 ==> SparseUnivariatePolynomial UP
  REC ==> Record(center:UP, equation:UP)

  Exports ==> with
    denomLODE: (L, RF) -> Union(UP, "failed")
      ++ denomLODE(op, g) returns a polynomial d such that
      ++ any rational solution of \spad{op y = g}
      ++ is of the form \spad{p/d} for some polynomial p, and
      ++ "failed", if the equation has no rational solution.
    denomLODE: (L, List RF) -> UP
      ++ denomLODE(op, [g1,...,gm]) returns a polynomial
      ++ d such that any rational solution of \spad{op y = c1 g1 + ... + cm gm}
      ++ is of the form \spad{p/d} for some polynomial p.
    indicialEquations: L -> List REC
      ++ indicialEquations op returns \spad{[[d1,e1],...,[dq,eq]]} where
      ++ the \spad{d_i}'s are the affine singularities of \spad{op},
      ++ and the \spad{e_i}'s are the indicial equations at each \spad{d_i}.
    indicialEquations: (L, UP) -> List REC
      ++ indicialEquations(op, p) returns \spad{[[d1,e1],...,[dq,eq]]} where
      ++ the \spad{d_i}'s are the affine singularities of \spad{op}
      ++ above the roots of \spad{p},
      ++ and the \spad{e_i}'s are the indicial equations at each \spad{d_i}.
    indicialEquation: (L, F) -> UP
      ++ indicialEquation(op, a) returns the indicial equation of \spad{op}
      ++ at \spad{a}.
    indicialEquations: LQ -> List REC
      ++ indicialEquations op returns \spad{[[d1,e1],...,[dq,eq]]} where
      ++ the \spad{d_i}'s are the affine singularities of \spad{op},
      ++ and the \spad{e_i}'s are the indicial equations at each \spad{d_i}.
    indicialEquations: (LQ, UP) -> List REC
      ++ indicialEquations(op, p) returns \spad{[[d1,e1],...,[dq,eq]]} where
      ++ the \spad{d_i}'s are the affine singularities of \spad{op}
      ++ above the roots of \spad{p},
      ++ and the \spad{e_i}'s are the indicial equations at each \spad{d_i}.
    indicialEquation: (LQ, F) -> UP
      ++ indicialEquation(op, a) returns the indicial equation of \spad{op}
      ++ at \spad{a}.
    splitDenominator: (LQ, List RF) -> Record(eq:L, rh:List RF)
      ++ splitDenominator(op, [g1,...,gm]) returns \spad{op0, [h1,...,hm]}
      ++ such that the equations \spad{op y = c1 g1 + ... + cm gm} and
      ++ \spad{op0 y = c1 h1 + ... + cm hm} have the same solutions.

  Implementation ==> add
    import BoundIntegerRoots(F, UP)
    import BalancedFactorisation(F, UP)
    import InnerCommonDenominator(UP, RF, List UP, List RF)
    import UnivariatePolynomialCategoryFunctions2(F, UP, UP, UP2)

    tau          : (UP, UP, UP, N) -> UP
    NPbound      : (UP, L, UP) -> N
    hdenom       : (L, UP, UP) -> UP
    denom0       : (Z, L, UP, UP, UP) -> UP
    indicialEq   : (UP, List N, List UP) -> UP
    separateZeros: (UP, UP) -> UP
    UPfact       : N -> UP
    UP2UP2       : UP -> UP2
    indeq        : (UP, L) -> UP
    NPmulambda   : (UP, L) -> Record(mu:Z, lambda:List N, func:List UP)

    diff := D()$L

    UP2UP2 p                    == map(#1::UP, p)
    indicialEquations(op:L)     == indicialEquations(op, leadingCoefficient op)
    indicialEquation(op:L, a:F) == indeq(monomial(1, 1) - a::UP, op)

    splitDenominator(op, lg) ==
      import RF
      cd := splitDenominator coefficients op
      f  := cd.den / gcd(cd.num)
      l:L := 0
      while op ~= 0 repeat
          l  := l + monomial(retract(f * leadingCoefficient op), degree op)
          op := reductum op
      [l, [f * g for g in lg]]

    tau(p, pp, q, n) ==
      ((pp ** n) * ((q exquo (p ** order(q, p)))::UP)) rem p

    indicialEquations(op:LQ) ==
      indicialEquations(splitDenominator(op, empty()).eq)

    indicialEquations(op:LQ, p:UP) ==
      indicialEquations(splitDenominator(op, empty()).eq, p)

    indicialEquation(op:LQ, a:F) ==
      indeq(monomial(1, 1) - a::UP, splitDenominator(op, empty()).eq)

-- returns z(z-1)...(z-(n-1))
    UPfact n ==
      zero? n => 1
      z := monomial(1, 1)$UP
      */[z - i::F::UP for i in 0..(n-1)::N]

    indicialEq(c, lamb, lf) ==
      cp := diff c
      cc := UP2UP2 c
      s:UP2 := 0
      for i in lamb for f in lf repeat
        s := s + (UPfact i) * UP2UP2 tau(c, cp, f, i)
      primitivePart resultant(cc, s)

    NPmulambda(c, l) ==
      lamb:List(N) := [d := degree l]
      lf:List(UP) := [a := leadingCoefficient l]
      mup := d::Z - order(a, c)
      while (l := reductum l) ~= 0 repeat
          a := leadingCoefficient l
          if (m := (d := degree l)::Z - order(a, c)) > mup then
              mup := m
              lamb := [d]
              lf := [a]
          else if (m = mup) then
              lamb := concat(d, lamb)
              lf := concat(a, lf)
      [mup, lamb, lf]

-- e = 0 means homogeneous equation
    NPbound(c, l, e) ==
      rec := NPmulambda(c, l)
      n := max(0, - integerBound indicialEq(c, rec.lambda, rec.func))
      zero? e => n::N
      max(n, order(e, c)::Z - rec.mu)::N

    hdenom(l, d, e) ==
      */[dd.factor ** NPbound(dd.factor, l, e)
                    for dd in factors balancedFactorisation(d, coefficients l)]

    denom0(n, l, d, e, h) ==
      hdenom(l, d, e) * */[hh.factor ** max(0, order(e, hh.factor) - n)::N
                                 for hh in factors balancedFactorisation(h, e)]

-- returns a polynomials whose zeros are the zeros of e which are not
-- zeros of d
    separateZeros(d, e) ==
      ((g := squareFreePart e) exquo gcd(g, squareFreePart d))::UP

    indeq(c, l) ==
      rec := NPmulambda(c, l)
      indicialEq(c, rec.lambda, rec.func)

    indicialEquations(op:L, p:UP) ==
      [[dd.factor, indeq(dd.factor, op)]
                   for dd in factors balancedFactorisation(p, coefficients op)]

-- cannot return "failed" in the homogeneous case
    denomLODE(l:L, g:RF) ==
      d := leadingCoefficient l
      zero? g => hdenom(l, d, 0)
      h := separateZeros(d, e := denom g)
      n := degree l
      (e exquo (h**(n + 1))) case "failed" => "failed"
      denom0(n, l, d, e, h)

    denomLODE(l:L, lg:List RF) ==
      empty? lg => denomLODE(l, 0)::UP
      d := leadingCoefficient l
      h := separateZeros(d, e := "lcm"/[denom g for g in lg])
      denom0(degree l, l, d, e, h)

@
\section{package UTSODETL UTSodetools}
<<package UTSODETL UTSodetools>>=
)abbrev package UTSODETL UTSodetools
++ Author: Manuel Bronstein
++ Date Created: 31 January 1994
++ Date Last Updated: 3 February 1994
++ Description:
++  \spad{RUTSodetools} provides tools to interface with the series
++   ODE solver when presented with linear ODEs.
UTSodetools(F, UP, L, UTS): Exports == Implementation where
  F  : Ring
  UP : UnivariatePolynomialCategory F
  L  : LinearOrdinaryDifferentialOperatorCategory UP
  UTS: UnivariateTaylorSeriesCategory F

  Exports ==> with
      UP2UTS:   UP -> UTS
          ++ UP2UTS(p) converts \spad{p} to a Taylor series.
      UTS2UP:   (UTS, NonNegativeInteger) -> UP
          ++ UTS2UP(s, n) converts the first \spad{n} terms of \spad{s}
          ++ to a univariate polynomial.
      LODO2FUN: L -> (List UTS -> UTS)
          ++ LODO2FUN(op) returns the function to pass to the series ODE
          ++ solver in order to solve \spad{op y = 0}.
      if F has IntegralDomain then
          RF2UTS: Fraction UP -> UTS
              ++ RF2UTS(f) converts \spad{f} to a Taylor series.

  Implementation ==> add
      fun: (Vector UTS, List UTS) -> UTS

      UP2UTS p ==
        q := p(monomial(1, 1) + center(0)::UP)
        +/[monomial(coefficient(q, i), i)$UTS for i in 0..degree q]

      UTS2UP(s, n) ==
        xmc     := monomial(1, 1)$UP - center(0)::UP
        xmcn:UP := 1
        ans:UP  := 0
        for i in 0..n repeat
            ans  := ans + coefficient(s, i) * xmcn
            xmcn := xmc * xmcn
        ans

      LODO2FUN op ==
          a := recip(UP2UTS(- leadingCoefficient op))::UTS
          n := (degree(op) - 1)::NonNegativeInteger
          v := [a * UP2UTS coefficient(op, i) for i in 0..n]$Vector(UTS)
          fun(v, #1)

      fun(v, l) ==
          ans:UTS := 0
          for b in l for i in 1.. repeat ans := ans + v.i * b
          ans

      if F has IntegralDomain then
          RF2UTS f == UP2UTS(numer f) * recip(UP2UTS denom f)::UTS

@
\section{package ODERAT RationalLODE}
<<package ODERAT RationalLODE>>=
)abbrev package ODERAT RationalLODE
++ Author: Manuel Bronstein
++ Date Created: 13 March 1991
++ Date Last Updated: 13 April 1994
++ Description:
++  \spad{RationalLODE} provides functions for in-field solutions of linear
++   ordinary differential equations, in the rational case.
RationalLODE(F, UP): Exports == Implementation where
  F  : Join(Field, CharacteristicZero, RetractableTo Integer,
                                       RetractableTo Fraction Integer)
  UP : UnivariatePolynomialCategory F

  N   ==> NonNegativeInteger
  Z   ==> Integer
  RF  ==> Fraction UP
  U   ==> Union(RF, "failed")
  V   ==> Vector F
  M   ==> Matrix F
  LODO ==> LinearOrdinaryDifferentialOperator1 RF
  LODO2==> LinearOrdinaryDifferentialOperator2(UP, RF)

  Exports ==> with
    ratDsolve: (LODO, RF) -> Record(particular: U, basis: List RF)
      ++ ratDsolve(op, g) returns \spad{["failed", []]} if the equation
      ++ \spad{op y = g} has no rational solution. Otherwise, it returns
      ++ \spad{[f, [y1,...,ym]]} where f is a particular rational solution
      ++ and the yi's form a basis for the rational solutions of the
      ++ homogeneous equation.
    ratDsolve: (LODO, List RF) -> Record(basis:List RF, mat:Matrix F)
      ++ ratDsolve(op, [g1,...,gm]) returns \spad{[[h1,...,hq], M]} such
      ++ that any rational solution of \spad{op y = c1 g1 + ... + cm gm}
      ++ is of the form \spad{d1 h1 + ... + dq hq} where
      ++ \spad{M [d1,...,dq,c1,...,cm] = 0}.
    ratDsolve: (LODO2, RF) -> Record(particular: U, basis: List RF)
      ++ ratDsolve(op, g) returns \spad{["failed", []]} if the equation
      ++ \spad{op y = g} has no rational solution. Otherwise, it returns
      ++ \spad{[f, [y1,...,ym]]} where f is a particular rational solution
      ++ and the yi's form a basis for the rational solutions of the
      ++ homogeneous equation.
    ratDsolve: (LODO2, List RF) -> Record(basis:List RF, mat:Matrix F)
      ++ ratDsolve(op, [g1,...,gm]) returns \spad{[[h1,...,hq], M]} such
      ++ that any rational solution of \spad{op y = c1 g1 + ... + cm gm}
      ++ is of the form \spad{d1 h1 + ... + dq hq} where
      ++ \spad{M [d1,...,dq,c1,...,cm] = 0}.
    indicialEquationAtInfinity: LODO -> UP
      ++ indicialEquationAtInfinity op returns the indicial equation of
      ++ \spad{op} at infinity.
    indicialEquationAtInfinity: LODO2 -> UP
      ++ indicialEquationAtInfinity op returns the indicial equation of
      ++ \spad{op} at infinity.

  Implementation ==> add
    import BoundIntegerRoots(F, UP)
    import RationalIntegration(F, UP)
    import PrimitiveRatDE(F, UP, LODO2, LODO)
    import LinearSystemMatrixPackage(F, V, V, M)
    import InnerCommonDenominator(UP, RF, List UP, List RF)

    nzero?             : V -> Boolean
    evenodd            : N -> F
    UPfact             : N -> UP
    infOrder           : RF -> Z
    infTau             : (UP, N) -> F
    infBound           : (LODO2, List RF) -> N
    regularPoint       : (LODO2, List RF) -> Z
    infIndicialEquation: (List N, List UP) -> UP
    makeDot            : (Vector F, List RF) -> RF
    unitlist           : (N, N) -> List F
    infMuLambda: LODO2 -> Record(mu:Z, lambda:List N, func:List UP)
    ratDsolve0: (LODO2, RF) -> Record(particular: U, basis: List RF)
    ratDsolve1: (LODO2, List RF) -> Record(basis:List RF, mat:Matrix F)
    candidates: (LODO2,List RF,UP) -> Record(basis:List RF,particular:List RF)

    dummy := new()$Symbol

    infOrder f == (degree denom f) - (degree numer f)
    evenodd n  == (even? n => 1; -1)

    ratDsolve1(op, lg) ==
      d := denomLODE(op, lg)
      rec := candidates(op, lg, d)
      l := concat([op q for q in rec.basis],
                  [op(rec.particular.i) - lg.i for i in 1..#(rec.particular)])
      sys1 := reducedSystem(matrix [l])@Matrix(UP)
      [rec.basis, reducedSystem sys1]

    ratDsolve0(op, g) ==
      zero? degree op => [inv(leadingCoefficient(op)::RF) * g, empty()]
      positive? minimumDegree op =>
        sol := ratDsolve0(monicRightDivide(op, monomial(1, 1)).quotient, g)
        b:List(RF) := [1]
        for f in sol.basis repeat
          if (uu := infieldint f) case RF then b := concat(uu::RF, b)
        sol.particular case "failed" => ["failed", b]
        [infieldint(sol.particular::RF), b]
      (u := denomLODE(op, g)) case "failed" => ["failed", empty()]
      rec := candidates(op, [g], u::UP)
      l := lb := lsol := empty()$List(RF)
      for q in rec.basis repeat
          if zero?(opq := op q) then lsol := concat(q, lsol)
          else (l := concat(opq, l); lb := concat(q, lb))
      h:RF := (zero? g => 0; first(rec.particular))
      empty? l =>
          zero? g => [0, lsol]
          [(g = op h => h; "failed"), lsol]
      m:M
      v:V
      if zero? g then
          m := reducedSystem(reducedSystem(matrix [l])@Matrix(UP))@M
          v := new(ncols m, 0)$V
      else
          sys1 := reducedSystem(matrix [l], vector [g - op h]
                               )@Record(mat: Matrix UP, vec: Vector UP)
          sys2 := reducedSystem(sys1.mat, sys1.vec)@Record(mat:M, vec:V)
          m := sys2.mat
          v := sys2.vec
      sol := solve(m, v)
      part:U :=
        zero? g => 0
        sol.particular case "failed" => "failed"
        makeDot(sol.particular::V, lb) + first(rec.particular)
      [part,
       concat!(lsol, [makeDot(v, lb) for v in sol.basis | nzero? v])]

    indicialEquationAtInfinity(op:LODO2) ==
      rec := infMuLambda op
      infIndicialEquation(rec.lambda, rec.func)

    indicialEquationAtInfinity(op:LODO) ==
      rec := splitDenominator(op, empty())
      indicialEquationAtInfinity(rec.eq)

    regularPoint(l, lg) ==
      a := leadingCoefficient(l) * commonDenominator lg
      coefficient(a, 0) ~= 0 => 0
      for i in 1.. repeat
        a(j := i::F) ~= 0 => return i
        a(-j) ~= 0 => return(-i)

    unitlist(i, q) ==
      v := new(q, 0)$Vector(F)
      v.i := 1
      members v

    candidates(op, lg, d) ==
      n := degree d + infBound(op, lg)
      m := regularPoint(op, lg)
      uts := UnivariateTaylorSeries(F, dummy, m::F)
      tools := UTSodetools(F, UP, LODO2, uts)
      solver := UnivariateTaylorSeriesODESolver(F, uts)
      dd := UP2UTS(d)$tools
      f := LODO2FUN(op)$tools
      q := degree op
      e := unitlist(1, q)
      hom := [UTS2UP(dd * ode(f, unitlist(i, q))$solver, n)$tools /$RF d
                   for i in 1..q]$List(RF)
      a1 := inv(leadingCoefficient(op)::RF)
      part := [UTS2UP(dd * ode(RF2UTS(a1 * g)$tools + f #1, e)$solver, n)$tools
                /$RF d for g in lg | g ~= 0]$List(RF)
      [hom, part]

    nzero? v ==
      for i in minIndex v .. maxIndex v repeat
        not zero? qelt(v, i) => return true
      false

-- returns z(z+1)...(z+(n-1))
    UPfact n ==
      zero? n => 1
      z := monomial(1, 1)$UP
      */[z + i::F::UP for i in 0..(n-1)::N]

    infMuLambda l ==
      lamb:List(N) := [d := degree l]
      lf:List(UP) := [a := leadingCoefficient l]
      mup := degree(a)::Z - d
      while (l := reductum l) ~= 0 repeat
          a := leadingCoefficient l
          if (m := degree(a)::Z - (d := degree l)) > mup then
            mup := m
            lamb := [d]
            lf := [a]
          else if (m = mup) then
            lamb := concat(d, lamb)
            lf := concat(a, lf)
      [mup, lamb, lf]

    infIndicialEquation(lambda, lf) ==
      ans:UP := 0
      for i in lambda for f in lf repeat
        ans := ans + evenodd i * leadingCoefficient f * UPfact i
      ans

    infBound(l, lg) ==
      rec := infMuLambda l
      n := min(- degree(l)::Z - 1,
               integerBound infIndicialEquation(rec.lambda, rec.func))
      while not(empty? lg) and zero? first lg repeat lg := rest lg
      empty? lg => (-n)::N
      m := infOrder first lg
      for g in rest lg repeat
        if not(zero? g) and (mm := infOrder g) < m then m := mm
      (-min(n, rec.mu - degree(leadingCoefficient l)::Z + m))::N

    makeDot(v, bas) ==
      ans:RF := 0
      for i in 1.. for b in bas repeat ans := ans + v.i::UP * b
      ans

    ratDsolve(op:LODO, g:RF) ==
      rec := splitDenominator(op, [g])
      ratDsolve0(rec.eq, first(rec.rh))

    ratDsolve(op:LODO, lg:List RF) ==
      rec := splitDenominator(op, lg)
      ratDsolve1(rec.eq, rec.rh)

    ratDsolve(op:LODO2, g:RF) ==
      unit?(c := content op) => ratDsolve0(op, g)
      ratDsolve0((op exquo c)::LODO2, inv(c::RF) * g)

    ratDsolve(op:LODO2, lg:List RF) ==
      unit?(c := content op) => ratDsolve1(op, lg)
      ratDsolve1((op exquo c)::LODO2, [inv(c::RF) * g for g in lg])

@
\section{package ODETOOLS ODETools}
<<package ODETOOLS ODETools>>=
)abbrev package ODETOOLS ODETools
++ Author: Manuel Bronstein
++ Date Created: 20 March 1991
++ Date Last Updated: 2 February 1994
++ Description:
++   \spad{ODETools} provides tools for the linear ODE solver.
ODETools(F, LODO): Exports == Implementation where
  N ==> NonNegativeInteger
  L ==> List F
  V ==> Vector F
  M ==> Matrix F

  F:    Field
  LODO: LinearOrdinaryDifferentialOperatorCategory F

  Exports ==> with
    wronskianMatrix: L -> M
      ++ wronskianMatrix([f1,...,fn]) returns the \spad{n x n} matrix m
      ++ whose i^th row is \spad{[f1^(i-1),...,fn^(i-1)]}.
    wronskianMatrix: (L, N) -> M
      ++ wronskianMatrix([f1,...,fn], q, D) returns the \spad{q x n} matrix m
      ++ whose i^th row is \spad{[f1^(i-1),...,fn^(i-1)]}.
    variationOfParameters: (LODO, F, L) -> Union(V, "failed")
      ++ variationOfParameters(op, g, [f1,...,fm])
      ++ returns \spad{[u1,...,um]} such that a particular solution of the
      ++ equation \spad{op y = g} is \spad{f1 int(u1) + ... + fm int(um)}
      ++ where \spad{[f1,...,fm]} are linearly independent and \spad{op(fi)=0}.
      ++ The value "failed" is returned if \spad{m < n} and no particular
      ++ solution is found.
    particularSolution: (LODO, F, L, F -> F) -> Union(F, "failed")
      ++ particularSolution(op, g, [f1,...,fm], I) returns a particular
      ++ solution h of the equation \spad{op y = g} where \spad{[f1,...,fm]}
      ++ are linearly independent and \spad{op(fi)=0}.
      ++ The value "failed" is returned if no particular solution is found.
      ++ Note: the method of variations of parameters is used.

  Implementation ==> add
    import LinearSystemMatrixPackage(F, V, V, M)

    diff := D()$LODO

    wronskianMatrix l == wronskianMatrix(l, #l)

    wronskianMatrix(l, q) ==
      v:V := vector l
      m:M := zero(q, #v)
      for i in minRowIndex m .. maxRowIndex m repeat
        setRow!(m, i, v)
        v := map!(diff #1, v)
      m

    variationOfParameters(op, g, b) ==
      empty? b => "failed"
      v:V := new(n := degree op, 0)
      qsetelt!(v, maxIndex v, g / leadingCoefficient op)
      particularSolution(wronskianMatrix(b, n), v)

    particularSolution(op, g, b, integration) ==
      zero? g => 0
      (sol := variationOfParameters(op, g, b)) case "failed" => "failed"
      ans:F := 0
      for f in b for i in minIndex(s := sol::V) .. repeat
        ans := ans + integration(qelt(s, i)) * f
      ans

@
\section{package ODEINT ODEIntegration}
<<package ODEINT ODEIntegration>>=
)abbrev package ODEINT ODEIntegration
++ Author: Manuel Bronstein
++ Date Created: 4 November 1991
++ Date Last Updated: 2 February 1994
++ Description:
++ \spadtype{ODEIntegration} provides an interface to the integrator.
++ This package is intended for use
++ by the differential equations solver but not at top-level.
ODEIntegration(R, F): Exports == Implementation where
  R: Join(EuclideanDomain, RetractableTo Integer,
          LinearlyExplicitRingOver Integer, CharacteristicZero)
  F: Join(AlgebraicallyClosedFunctionSpace R, TranscendentalFunctionCategory,
                                              PrimitiveFunctionCategory)

  Q   ==> Fraction Integer
  UQ  ==> Union(Q, "failed")
  SY  ==> Symbol
  K   ==> Kernel F
  P   ==> SparseMultivariatePolynomial(R, K)
  REC ==> Record(coef:Q, logand:F)

  Exports ==> with
    int   : (F, SY) -> F
      ++ int(f, x) returns the integral of f with respect to x.
    expint: (F, SY) -> F
      ++ expint(f, x) returns e^{the integral of f with respect to x}.
    diff  : SY -> (F -> F)
      ++ diff(x) returns the derivation with respect to x.

  Implementation ==> add
    import FunctionSpaceIntegration(R, F)
    import ElementaryFunctionStructurePackage(R, F)

    isQ   : List F -> UQ
    isQlog: F -> Union(REC, "failed")
    mkprod: List REC -> F

    diff x == differentiate(#1, x)

-- This is the integration function to be used for quadratures
    int(f, x) ==
      (u := integrate(f, x)) case F => u::F
      first(u::List(F))

-- mkprod([q1, f1],...,[qn,fn]) returns */(fi^qi) but groups the
-- qi having the same denominator together
    mkprod l ==
      empty? l => 1
      rec := first l
      d := denom(rec.coef)
      ll := select(denom(#1.coef) = d, l)
      nthRoot(*/[r.logand ** numer(r.coef) for r in ll], d) *
        mkprod setDifference(l, ll)

-- computes exp(int(f,x)) in a non-naive way
    expint(f, x) ==
      a := int(f, x)
      (u := validExponential(tower a, a, x)) case F => u::F
      da := denom a
      l :=
        (v := isPlus(na := numer a)) case List(P) => v::List(P)
        [na]
      exponent:P := 0
      lrec:List(REC) := empty()
      for term in l repeat
        if (w := isQlog(term / da)) case REC then
          lrec := concat(w::REC, lrec)
        else
          exponent := exponent + term
      mkprod(lrec) * exp(exponent / da)

-- checks if all the elements of l are rational numbers, returns their product
    isQ l ==
      prod:Q := 1
      for x in l repeat
        (u := retractIfCan(x)@UQ) case "failed" => return "failed"
        prod := prod * u::Q
      prod

-- checks if a non-sum expr is of the form c * log(g) for a rational number c
    isQlog f ==
      is?(f, 'log) => [1, first argument(retract(f)@K)]
      (v := isTimes f) case List(F) and (#(l := v::List(F)) <= 3) =>
          l := reverse! sort!(before?,l)
          is?(first l, 'log) and ((u := isQ rest l) case Q) =>
              [u::Q, first argument(retract(first(l))@K)]
          "failed"
      "failed"

@
\section{package ODECONST ConstantLODE}
<<package ODECONST ConstantLODE>>=
)abbrev package ODECONST ConstantLODE
++ Author: Manuel Bronstein
++ Date Created: 18 March 1991
++ Date Last Updated: 3 February 1994
++ Description: Solution of linear ordinary differential equations, constant coefficient case.
ConstantLODE(R, F, L): Exports == Implementation where
  R: Join(EuclideanDomain, RetractableTo Integer,
          LinearlyExplicitRingOver Integer, CharacteristicZero)
  F: Join(AlgebraicallyClosedFunctionSpace R,
          TranscendentalFunctionCategory, PrimitiveFunctionCategory)
  L: LinearOrdinaryDifferentialOperatorCategory F

  Z   ==> Integer
  SY  ==> Symbol
  K   ==> Kernel F
  V   ==> Vector F
  M   ==> Matrix F
  SUP ==> SparseUnivariatePolynomial F

  Exports ==> with
    constDsolve: (L, F, SY) -> Record(particular:F, basis:List F)
      ++ constDsolve(op, g, x) returns \spad{[f, [y1,...,ym]]}
      ++ where f is a particular solution of the equation \spad{op y = g},
      ++ and the \spad{yi}'s form a basis for the solutions of \spad{op y = 0}.

  Implementation ==> add
    import ODETools(F, L)
    import ODEIntegration(R, F)
    import ElementaryFunctionSign(R, F)
    import AlgebraicManipulations(R, F)
    import FunctionSpaceIntegration(R, F)
    import FunctionSpaceUnivariatePolynomialFactor(R, F, SUP)

    homoBasis: (L, F) -> List F
    quadSol  : (SUP, F) -> List F
    basisSqfr: (SUP, F) -> List F
    basisSol : (SUP, Z, F) -> List F

    constDsolve(op, g, x) ==
      b := homoBasis(op, x::F)
      [particularSolution(op, g, b, int(#1, x))::F, b]

    homoBasis(op, x) ==
      p:SUP := 0
      while op ~= 0 repeat
          p  := p + monomial(leadingCoefficient op, degree op)
          op := reductum op
      b:List(F) := empty()
      for ff in factors ffactor p repeat
        b := concat!(b, basisSol(ff.factor, dec(ff.exponent), x))
      b

    basisSol(p, n, x) ==
      l := basisSqfr(p, x)
      zero? n => l
      ll := copy l
      xn := x::F
      for i in 1..n repeat
        l := concat!(l, [xn * f for f in ll])
        xn := x * xn
      l

    basisSqfr(p, x) ==
      one?(d := degree p) =>
        [exp(- coefficient(p, 0) * x / leadingCoefficient p)]
      d = 2 => quadSol(p, x)
      [exp(a * x) for a in rootsOf p]

    quadSol(p, x) ==
      (u := sign(delta := (b := coefficient(p, 1))**2 - 4 *
        (a := leadingCoefficient p) * (c := coefficient(p, 0))))
          case Z and negative?(u::Z) =>
            y := x / (2 * a)
            r := - b * y
            i := rootSimp(sqrt(-delta)) * y
            [exp(r) * cos(i), exp(r) * sin(i)]
      [exp(a * x) for a in zerosOf p]

@
\section{License}
<<license>>=
--Copyright (c) 1991-2002, The Numerical ALgorithms Group Ltd.
--All rights reserved.
--
--Redistribution and use in source and binary forms, with or without
--modification, are permitted provided that the following conditions are
--met:
--
--    - Redistributions of source code must retain the above copyright
--      notice, this list of conditions and the following disclaimer.
--
--    - Redistributions in binary form must reproduce the above copyright
--      notice, this list of conditions and the following disclaimer in
--      the documentation and/or other materials provided with the
--      distribution.
--
--    - Neither the name of The Numerical ALgorithms Group Ltd. nor the
--      names of its contributors may be used to endorse or promote products
--      derived from this software without specific prior written permission.
--
--THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS
--IS" AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED
--TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A
--PARTICULAR PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER
--OR CONTRIBUTORS BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL,
--EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO,
--PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR
--PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF
--LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING
--NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF THIS
--SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
@
<<*>>=
<<license>>

-- Compile order for the differential equation solver:
-- oderf.spad  odealg.spad  nlode.spad  nlinsol.spad  riccati.spad  odeef.spad

<<package BALFACT BalancedFactorisation>>
<<package BOUNDZRO BoundIntegerRoots>>
<<package ODEPRIM PrimitiveRatDE>>
<<package UTSODETL UTSodetools>>
<<package ODERAT RationalLODE>>
<<package ODETOOLS ODETools>>
<<package ODEINT ODEIntegration>>
<<package ODECONST ConstantLODE>>
@
\eject
\begin{thebibliography}{99}
\bibitem{1} nothing
\end{thebibliography}
\end{document}