-
Notifications
You must be signed in to change notification settings - Fork 0
/
non-linear.lisp
972 lines (868 loc) · 40.8 KB
/
non-linear.lisp
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
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
; ACL2 Version 8.3 -- A Computational Logic for Applicative Common Lisp
; Copyright (C) 2020, Regents of the University of Texas
; This version of ACL2 is a descendent of ACL2 Version 1.9, Copyright
; (C) 1997 Computational Logic, Inc. See the documentation topic NOTE-2-0.
; This program is free software; you can redistribute it and/or modify
; it under the terms of the LICENSE file distributed with ACL2.
; This program is distributed in the hope that it will be useful,
; but WITHOUT ANY WARRANTY; without even the implied warranty of
; MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
; LICENSE for more details.
; Written by: Matt Kaufmann and J Strother Moore
; email: [email protected] and [email protected]
; Department of Computer Science
; University of Texas at Austin
; Austin, TX 78712 U.S.A.
(in-package "ACL2")
;=================================================================
; We here begin the support functions for non-linear arithmetic.
;=================================================================
(defun cleanse-type-alist (type-alist pt)
; This function removes equality facts from the type-alist in
; preparation for the call to rewrite-linear-term-lst in
; add-terms-and-lemmas.
; In order to see why we need this function now, redefine
; cleanse-type-alist to be the identity and trace
; rewrite-linear-term-lst and linearize in the (failed) proof
; of the following lemma:
; (defthm uniqueness-of-+-inverses-lemma
; (implies (and (acl2-numberp x)
; (acl2-numberp y)
; (equal (+ x y)
; 0))
; (equal (- x) y))
; :rule-classes nil)
(cond ((null type-alist)
nil)
((to-be-ignoredp (cddar type-alist) pt)
(cleanse-type-alist (cdr type-alist) pt))
(t
(cons (car type-alist)
(cleanse-type-alist (cdr type-alist) pt)))))
(defun var-with-divisionp (var)
; We test whether var is of the form
; 1. (/ x) or (expt (/ x) n), or [don't change this line without
; 2. (expt x c) or (expt x (* c y)) seeing the Warning below]
; where c is a negative constant.
; Warning: If you change this code, search for other instances of
; ``1. (/ x) or (expt (/ x) n)'' and adjust those comments
; appropriately.
; Warning: Keep this function in sync with invert-var.
(cond ((eq (fn-symb var) 'EXPT)
(let ((base (fargn var 1))
(exponent (fargn var 2)))
(or (and (eq (fn-symb base) 'UNARY-/)
(not (eq (fn-symb (fargn base 1)) 'BINARY-+)))
(and (not (eq (fn-symb base) 'BINARY-+))
(quotep exponent)
(integerp (unquote exponent))
(< (unquote exponent) 0))
(and (not (eq (fn-symb base) 'BINARY-+))
(eq (fn-symb exponent) 'BINARY-*)
(quotep (fargn exponent 1))
(integerp (unquote (fargn exponent 1)))
(< (unquote (fargn exponent 1)) 0)))))
(t
(and (eq (fn-symb var) 'UNARY-/)
(not (eq (fn-symb (fargn var 1)) 'BINARY-+))))))
(defun varify (x)
; We ensure that x is a legitimate pot-label. See invert-var from whence
; this is called.
(cond ((quotep x)
(er hard 'varify
"This should not have happened. The supposed ~
variable, ~x0, is instead a constant."
x))
((equal (fn-symb x) 'BINARY-+)
;;; We have to pick one.
(if (quotep (fargn x 1))
(varify (fargn x 2))
(varify (fargn x 1))))
((and (equal (fn-symb x) 'BINARY-*)
(quotep (fargn x 1)))
(varify (fargn x 2)))
(t
x)))
(defun varify! (x)
(let ((temp (varify x)))
(if (good-pot-varp temp)
temp
(er hard 'varify!
"Varify! is supposed to return a good-pot-varp, but ~
returned ~x0 on ~x1."
temp x))))
(defun varify!-lst1 (lst acc)
(if (null lst)
acc
(varify!-lst1 (cdr lst) (cons (varify! (car lst)) acc))))
(defun varify!-lst (lst)
; This is used in expanded-new-vars-in-pot-lst, and we want to
; reverse the list. Thus the use of an accumulator.
(varify!-lst1 lst nil))
(defun invert-var (var)
; Var is an arithmetic ACL2 term. We return a term suitable for use
; as an unknown in a poly, but that's all we guarantee. The idea is
; that the term is ``relevant'' to the non-linear properties of var
; and we try to return the multiplicative inverse. We expect to go
; find additional polys about this term.
(cond ((eq (fn-symb var) 'EXPT)
(let ((base (fargn var 1))
(exponent (fargn var 2)))
(cond ((eql exponent ''-1)
(varify! base))
((eq (fn-symb base) 'UNARY-/)
(fcons-term* 'EXPT (fargn base 1) exponent))
((eq (fn-symb exponent) 'UNARY--)
(fcons-term* 'EXPT base (fargn exponent 1)))
((and (quotep exponent)
(integerp (unquote exponent))
(< (unquote exponent) 0))
(fcons-term* 'EXPT base (kwote (- (unquote exponent)))))
((and (eq (fn-symb exponent) 'BINARY-*)
(quotep (fargn exponent 1))
(integerp (unquote (fargn exponent 1)))
(< (unquote (fargn exponent 1)) 0))
(fcons-term* 'EXPT
base
(cons-term
'BINARY-*
(list
(kwote (- (unquote (fargn exponent 1))))
(fargn exponent 2)))))
(t
(fcons-term* 'EXPT
(cons-term 'UNARY-/ (list base))
exponent)))))
((eq (fn-symb var) 'UNARY-/)
(varify! (fargn var 1)))
(t
(cons-term 'UNARY-/ (list var)))))
(defun part-of1 (var1 var2)
; NOTE: Note that we are implicitly assuming that var2 is right
; associated. This should be taken care of some day. Perhaps what I
; should do is take the fringe of both vars, and do a simple set
; difference.
(cond ((or (variablep var2)
(fquotep var2))
(if (equal var1 var2)
*1*
nil))
((eq (ffn-symb var2) 'BINARY-*)
(let ((arg1 (fargn var2 1))
(arg2 (fargn var2 2)))
(if (equal var1 arg1)
arg2
(let ((new-var2 (part-of1 var1 arg2)))
(cond ((null new-var2)
nil)
((equal new-var2 ''1)
arg1)
(t
(cons-term 'BINARY-*
(list arg1
new-var2))))))))
(t
(if (equal var1 var2)
*1*
nil))))
(defun part-of (var1 var2)
; We ask whether the factors of var1 are a subset of the factors of
; var2. If so, we return a naive calculation of (/ var2 var1).
(cond ((or (variablep var1)
(fquotep var1))
(part-of1 var1 var2))
((eq (ffn-symb var1) 'BINARY-*)
(let ((new-var2 (part-of (fargn var1 1) var2)))
(cond (new-var2
(part-of (fargn var1 2) new-var2))
(t
nil))))
(t
(part-of1 var1 var2))))
(defun product-already-triedp (var-list products-already-tried)
; Var-list is a list of ACL2 terms, products-already-tried is a
; list of lists of ACL2 terms. If (a permutation of) var-list
; appears in products-already-tried, we return t. Otherwise, we
; return nil.
(cond ((null products-already-tried)
nil)
(t
(or (equal var-list (car products-already-tried))
(product-already-triedp var-list
(cdr products-already-tried))))))
(defun too-many-polysp (var-lst pot-lst counter)
; Var-list is a list of pot-labels from pot-lst, and counter is initially
; 1. We we are about to multiply the polys from the pots in var-lst,
; we first check whether doing so would generate too many polys.
; Note: This function has a magic number, 20, which probably should be
; settable.
(cond
((null var-lst)
(< 20 counter))
(t
(too-many-polysp (cdr var-lst)
pot-lst
(* counter
(length (polys-with-var1 (car var-lst) pot-lst)))))))
(defun expanded-new-vars-in-pot-lst2 (new-var vars-to-skip vars-to-return)
; We explore the term new-var and accumulate into vars-to-skip and
; vars-to-return all the tips of the BINARY-* tree and the inverses of
; vars with division.
(cond
((or (member-equal new-var vars-to-skip)
(quotep new-var)
(eq (fn-symb new-var) 'BINARY-+))
(mv vars-to-skip vars-to-return))
((eq (fn-symb new-var) 'BINARY-*)
(let ((new-factor (fargn new-var 1)))
(if (or (member-equal new-factor vars-to-skip)
(quotep new-factor)
(eq (fn-symb new-factor) 'BINARY-+))
(expanded-new-vars-in-pot-lst2 (fargn new-var 2)
vars-to-skip
vars-to-return)
(expanded-new-vars-in-pot-lst2 (fargn new-var 2)
(cons new-factor vars-to-skip)
(cons new-factor vars-to-return)))))
((var-with-divisionp new-var)
(let ((inverted-var (invert-var new-var)))
(if (member-equal inverted-var vars-to-skip)
(mv (cons new-var vars-to-skip)
(cons new-var vars-to-return))
(mv (cons new-var (cons inverted-var vars-to-skip))
(cons new-var (cons inverted-var vars-to-return))))))
(t
(mv (cons new-var vars-to-skip)
(cons new-var vars-to-return)))))
(defun expanded-new-vars-in-pot-lst1 (new-pot-lst vars-to-skip
vars-to-return)
; We explore all the terms in new-pot-lst and collect all the ``new''
; vars (relative to vars-to-skip) and the inverses of the vars with
; division. We also collect all the factors of the vars we collect
; above, and their inverses. We accumulate onto both vars-to-skip and
; vars-to-return, but only return the latter.
(if (null new-pot-lst)
vars-to-return
(let ((new-var (access linear-pot (car new-pot-lst) :var)))
(cond
((member-equal new-var vars-to-skip)
(expanded-new-vars-in-pot-lst1 (cdr new-pot-lst)
vars-to-skip
vars-to-return))
((var-with-divisionp new-var)
(let* ((inverse-var (invert-var new-var))
; We keep vars-to-skip a superset of vars-to-return.
(new-vars-to-skip (if (member-equal inverse-var vars-to-skip)
(cons new-var vars-to-skip)
(cons new-var (cons inverse-var
vars-to-skip))))
(new-vars-to-return (if (member-equal inverse-var vars-to-skip)
(cons new-var vars-to-return)
(cons new-var (cons inverse-var
vars-to-return)))))
(expanded-new-vars-in-pot-lst1 (cdr new-pot-lst)
new-vars-to-skip
new-vars-to-return)))
((eq (fn-symb new-var) 'BINARY-*)
(mv-let (new-vars-to-skip new-vars-to-return)
(expanded-new-vars-in-pot-lst2 new-var
vars-to-skip
vars-to-return)
(expanded-new-vars-in-pot-lst1 (cdr new-pot-lst)
(cons new-var new-vars-to-skip)
(cons new-var new-vars-to-return))))
(t
(expanded-new-vars-in-pot-lst1 (cdr new-pot-lst)
(cons new-var vars-to-skip)
(cons new-var vars-to-return)))))))
(defun expanded-new-vars-in-pot-lst (new-pot-lst old-pot-lst)
; This is a variant of new-vars-in-pot-lst. See the comments there.
; Here, if a new var is a product, we recursively add all its individual
; factors to the list of new vars as well as the product itself.
; E. g., if a new var is (* x (foo y) (bar z)), we add
; x, (foo y), (bar z), and (* x (foo y) (bar z)) to the new var list.
; In addition, if a var or one of its factors is of the form
; 1. (/ x) or (expt (/ x) n) or
; 2. (expt x c) or (expt x (* c y)), where c is a negative integer,
; we expand the multiplicative inverse.
; We need to generate a list of vars to skip. We expand the vars in
; old-pot-lst.
(let ((vars-to-skip (expanded-new-vars-in-pot-lst1
old-pot-lst
nil ; vars-to-skip
nil)))
(varify!-lst
(expanded-new-vars-in-pot-lst1 new-pot-lst
vars-to-skip
nil))))
(defun extract-bounds (bounds-polys)
; Bounds-polys is a list of bounds-polys as returned by bounds-polys-with-var.
; It is either nil meaning no bounds were found, a list of one element
; which is either an upper or lower bound poly, or a list of length two
; consisting of an upper and lower bound poly. We return six items ---
; any of which may be nil if it was not found --- the lower bound, its
; relation, its ttree, an upper bound, its relation, and its ttree.
(cond ((null bounds-polys)
(mv nil nil nil nil nil nil))
((null (cdr bounds-polys))
;; We have only a lower, or upper bound.
(if (< (first-coefficient (car bounds-polys)) 0)
;; ((var < c)), we have only an upper bound.
(mv nil nil nil
(access poly (car bounds-polys) :constant)
(access poly (car bounds-polys) :relation)
(access poly (car bounds-polys) :ttree))
;; ((c < var)), we have only a lower bound.
(mv (- (access poly (car bounds-polys) :constant))
(access poly (car bounds-polys) :relation)
(access poly (car bounds-polys) :ttree)
nil nil nil)))
(t
;; We have both a lower and upper bound.
(if (< (first-coefficient (car bounds-polys)) 0)
;; ((var < c) (c < var)), upper bound lower bound.
(mv (- (access poly (cadr bounds-polys) :constant))
(access poly (cadr bounds-polys) :relation)
(access poly (cadr bounds-polys) :ttree)
(access poly (car bounds-polys) :constant)
(access poly (car bounds-polys) :relation)
(access poly (car bounds-polys) :ttree))
;; ((c < var) (var < c)), lower bound upper bound.
(mv (- (access poly (car bounds-polys) :constant))
(access poly (car bounds-polys) :relation)
(access poly (car bounds-polys) :ttree)
(access poly (cadr bounds-polys) :constant)
(access poly (cadr bounds-polys) :relation)
(access poly (cadr bounds-polys) :ttree))))))
(defun good-bounds-in-pot (var pot-lst pt)
; Is bds good enough for deal-with-division? That is, can we use
; the information in pot-lst to bound var either away from zero or
; at zero?
; This information is needed in deal-with-division where we may,
; for instance, multiply x and (/ x). If we know that x is non-zero,
; their product will be equal to one. If we know that x is zero, their
; product is equal to zero.
(let ((bounds-polys (bounds-polys-with-var var
pot-lst
pt)))
(mv-let (var-lbd var-lbd-rel var-lbd-ttree
var-ubd var-ubd-rel var-ubd-ttree)
(extract-bounds bounds-polys)
(declare (ignore var-lbd-ttree var-ubd-ttree))
(or (and (eql var-lbd 0)
(eql var-ubd 0))
(and var-lbd
(< 0 var-lbd))
(and (eql var-lbd 0)
(eq var-lbd-rel '<))
(and var-ubd
(< var-ubd 0))
(and (eql var-ubd 0)
(eq var-ubd-rel '<))))))
(defun inverse-polys (var inv-var pot-lst ttree pt)
; Var and inv-var are as in add-inverse-polys. Ttree is the ttree
; from the call to type-set in inverse-polys.
; Bounds-polys-for-var extracts any bounds polys from pot-lst.
; Extract-bounds deconstructs any bounds polys found. This
; information is then used to try to tighten up our knowledge about
; both var and inv-var. We return a list of bounds polys constructed
; using this information.
; A couple of simple examples will help to understand what we are doing:
; 1. If we can determine that (< 4 x), we can add both (< 0 (/ x)) and
; (< (/ x) 1/4).
; 2. If we can determine that (< 0 x) and (< x 4), we can add
; (< 0 (/ x)) and (< (/ x) 1/4).
; 3. If we can only determine that (< -2 x), we cannot add anything about
; (/ x) to the pot-lst.
; We handle the symmetric cases for negative x.
(if (and (good-pot-varp var)
(good-pot-varp inv-var))
(let ((bounds-polys-for-var
(bounds-polys-with-var var pot-lst pt))
(bounds-polys-for-inv-var
(bounds-polys-with-var inv-var pot-lst pt)))
(mv-let (var-lbd var-lbd-rel var-lbd-ttree
var-ubd var-ubd-rel var-ubd-ttree)
(extract-bounds bounds-polys-for-var)
(mv-let (inv-var-lbd inv-var-lbd-rel inv-var-lbd-ttree
inv-var-ubd inv-var-ubd-rel inv-var-ubd-ttree)
(extract-bounds bounds-polys-for-inv-var)
(cond
((and (or (eql var-lbd 0)
(eql inv-var-lbd 0))
(or (eql var-ubd 0)
(eql inv-var-ubd 0)))
; Assume that all four relations are <=. That is a weaker assumption
; than whatever is really the case. From that assumption, we conclude
; that var and inv-var are 0. We make sure that this is recorded in
; the pot-lst. If the actual case is that one of the relations is a
; strict <, then the case is contradictory and we can add any other
; polys we want, including these. Note that at least two of the four
; polys we are about to add are already in the pot-lst.
(list
;; 0 <= var
(add-linear-terms :rhs var
(base-poly (cons-tag-trees
ttree
(cons-tag-trees var-lbd-ttree
inv-var-lbd-ttree))
'<=
t
nil))
;; var <= 0
(add-linear-terms :lhs var
(base-poly (cons-tag-trees
ttree
(cons-tag-trees var-ubd-ttree
inv-var-ubd-ttree))
'<=
t
nil))
;; 0 <= inv-var
(add-linear-terms :rhs inv-var
(base-poly (cons-tag-trees
ttree
(cons-tag-trees var-lbd-ttree
inv-var-lbd-ttree))
'<=
t
nil))
;; inv-var <= 0
(add-linear-terms :lhs inv-var
(base-poly (cons-tag-trees
ttree
(cons-tag-trees var-ubd-ttree
inv-var-ubd-ttree))
'<=
t
nil))))
((or (and var-lbd
(< 0 var-lbd))
(and inv-var-lbd
(< 0 inv-var-lbd)))
; We try to gather bounds polys in four stages --- a lower bound for inv-var,
; a lower bound for var, an upper bound for inv-var, and an upper bound
; for var.
(let* ((ttree1 (cons-tag-trees ttree
(cons-tag-trees var-lbd-ttree
inv-var-lbd-ttree)))
(bounds-polys1
(cond ((and var-ubd
(not (eql var-ubd 0))
(or (null inv-var-lbd)
(< inv-var-lbd (/ var-ubd))))
(list
;; (/ var-ubd) [<,<=] inv-var
(add-linear-terms :lhs (kwote (/ var-ubd))
:rhs inv-var
(base-poly (cons-tag-trees
ttree1
var-ubd-ttree)
var-ubd-rel
t
nil))))
((null inv-var-lbd)
(list
;; 0 < inv-var
(add-linear-terms :rhs inv-var
(base-poly ttree1
'<
t
nil))))
(t
nil)))
(bounds-polys2
(cond ((and inv-var-ubd
(not (eql inv-var-ubd 0))
(or (null var-lbd)
(< var-lbd (/ inv-var-ubd))))
(cons
;; (/ inv-var-ubd) [<,<=] var
(add-linear-terms :lhs (kwote (/ inv-var-ubd))
:rhs var
(base-poly (cons-tag-trees
ttree1
inv-var-ubd-ttree)
inv-var-ubd-rel
t
nil))
bounds-polys1))
((null var-lbd)
;; 0 < var
(cons
(add-linear-terms :rhs var
(base-poly ttree1
'<
t
nil))
bounds-polys1))
(t
bounds-polys1)))
(bounds-polys3
(cond ((and var-lbd
(not (eql var-lbd 0))
(or (null inv-var-ubd)
(< (/ var-lbd) inv-var-ubd)))
(cons
;; inv-var [<,<=] (/ var-lbd)
(add-linear-terms :lhs inv-var
:rhs (kwote (/ var-lbd))
(base-poly ttree1
var-lbd-rel
t
nil))
bounds-polys2))
(t
bounds-polys2)))
(bounds-polys4
(cond ((and inv-var-lbd
(not (eql inv-var-lbd 0))
(or (null var-ubd)
(< (/ inv-var-lbd) var-ubd)))
(cons
;; var [<,<=] (/ inv-var-lbd)
(add-linear-terms :lhs var
:rhs (kwote (/ inv-var-lbd))
(base-poly ttree1
inv-var-lbd-rel
t
nil))
bounds-polys3))
(t
bounds-polys3))))
bounds-polys4))
((or (and var-ubd
(< var-ubd 0))
(and inv-var-ubd
(< inv-var-ubd 0)))
; We try to gather bounds polys in four stages --- a upper bound for inv-var,
; a upper bound for var, an lower bound for inv-var, and an lower bound
; for var.
(let* ((ttree1 (cons-tag-trees ttree
(cons-tag-trees var-ubd-ttree
inv-var-ubd-ttree)))
(bounds-polys1
(cond ((and var-lbd
(not (eql var-lbd 0))
(or (null inv-var-ubd)
(< (/ var-lbd) inv-var-ubd)))
(list
;; inv-var [<,<=] (/ var-lbd)
(add-linear-terms :lhs inv-var
:rhs (kwote (/ var-lbd))
(base-poly (cons-tag-trees
ttree1
var-lbd-ttree)
var-lbd-rel
t
nil))))
((null inv-var-ubd)
(list
;; inv-var < 0
(add-linear-terms :lhs inv-var
(base-poly ttree1
'<
t
nil))))
(t
nil)))
(bounds-polys2
(cond ((and inv-var-lbd
(not (eql inv-var-lbd 0))
(or (null var-ubd)
(< (/ inv-var-lbd) var-ubd)))
(cons
;; var [<,<=] (/ inv-var-lbd)
(add-linear-terms :lhs var
:rhs (kwote (/ inv-var-lbd))
(base-poly (cons-tag-trees
ttree1
inv-var-lbd-ttree)
var-lbd-rel
t
nil))
bounds-polys1))
((null var-ubd)
;; var < 0
(cons
(add-linear-terms :lhs var
(base-poly ttree1
'<
t
nil))
bounds-polys1))
(t
bounds-polys1)))
(bounds-polys3
(cond ((and var-ubd
(not (eql var-ubd 0))
(or (null inv-var-lbd)
(< inv-var-lbd (/ var-ubd))))
(cons
;; (/ var-ubd) [<,<=] inv-var
(add-linear-terms :lhs (kwote (/ var-ubd))
:rhs inv-var
(base-poly ttree1
var-ubd-rel
t
nil))
bounds-polys2))
(t
bounds-polys2)))
(bounds-polys4
(cond ((and inv-var-ubd
(not (eql inv-var-ubd 0))
(or (null var-lbd)
(< var-lbd (/ inv-var-ubd))))
(cons
;; (/ inv-var-ubd) [<,<=] var
(add-linear-terms :lhs (kwote (/ inv-var-ubd))
:rhs var
(base-poly ttree1
inv-var-ubd-rel
t
nil))
bounds-polys3))
(t
bounds-polys3))))
bounds-polys4))
((and (eql var-lbd 0)
(eq var-lbd-rel '<))
;; 0 < inv-var
(list
(add-linear-terms :rhs inv-var
(base-poly (cons-tag-trees
ttree
var-lbd-ttree)
'<
t
nil))))
((and (eql inv-var-lbd 0)
(eq inv-var-lbd-rel '<))
;; 0 < var
(list
(add-linear-terms :rhs var
(base-poly (cons-tag-trees
ttree
inv-var-lbd-ttree)
'<
t
nil))))
((and (eql var-ubd 0)
(eq var-ubd-rel '<))
;; inv-var < 0
(list
(add-linear-terms :lhs inv-var
(base-poly (cons-tag-trees
ttree
var-ubd-ttree)
'<
t
nil))))
((and (eql inv-var-ubd 0)
(eq inv-var-ubd-rel '<))
;; var < 0
(list
(add-linear-terms :lhs var
(base-poly (cons-tag-trees
ttree
inv-var-ubd-ttree)
'<
t
nil))))
(t
nil)))))
(er hard 'inverse-polys
"A presumptive pot-label, ~x0, has turned out to be illegitimate. ~
If possible, please send a script reproducing this error ~
to the authors of ACL2."
(if (good-pot-varp var)
inv-var
var))))
(defun add-inverse-polys (var
type-alist wrld
simplify-clause-pot-lst
force-flg ens pt)
; If var is of the form
; 1. (/ x) or (expt (/ x) n) and x is rational, or
; 2. (expt x c) or (expt x (* c y)), where c is a negative integer
; and x is rational,
; we try to find bounds for either var or its multiplicative inverse and
; then use any bounds found to bound the other. The work of gathering
; the polys is done in inverse-polys.
(if (var-with-divisionp var)
(let ((inverted-var (invert-var var)))
(mv-let (base-ts base-ttree)
(type-set inverted-var
force-flg
nil ; dwp
type-alist
ens
wrld
nil ; ttree
nil ; pot-lst
nil) ; pt
(if (ts-real/rationalp base-ts)
(let ((inverse-polys
(inverse-polys var
inverted-var
simplify-clause-pot-lst
base-ttree
pt)))
(add-polys inverse-polys
simplify-clause-pot-lst
pt
t ; nonlinearp hint
type-alist
ens
force-flg
wrld))
(mv nil simplify-clause-pot-lst))))
(mv nil simplify-clause-pot-lst)))
(defun add-polys-from-type-set (var pot-lst type-alist
pt force-flg ens wrld)
; At the time of this writing, this function is called only from
; add-polys-and-lemmas3. When doing non-linear arithmetic, we
; have found this extra information gathering useful.
; Warning: This function should not be used with any terms that are
; not legitimate pot-vars. See the definition of good-pot-varp.
; Assuming that term is a legitimate pot-label --- meets all the
; invariants --- we do not have to normalize any of the polys below.
; It would, however, not be very expensive to wrap the below in a call
; to normalize-poly-lst.
(if (good-pot-varp var)
(add-polys (polys-from-type-set var
force-flg
nil ; dwp
type-alist
ens
wrld
nil) ; ttree
pot-lst
pt
t ; nonlinearp hint
type-alist
ens
force-flg
wrld)
(mv (er hard 'add-polys-from-type-set
"A presumptive pot-label, ~x0, has turned out to be illegitimate. ~
If possible, please send a script reproducing this error ~
to the authors of ACL2."
var)
nil)))
(defun length-of-shortest-polys-with-var (poly-lst pt n)
; Poly-lst is a list of polys, pt is a parent tree of polys to be
; ignored, and n is the length of the shortest alist found so far, or
; t if we haven't found any yet. We cdr down pot-lst looking for a
; poly whose alist is shorter than n, and return the length of the
; shortest alist found.
(cond ((null poly-lst)
n)
((and (or (eq n t)
(< (length (access poly (car poly-lst) :alist)) n))
(not (ignore-polyp (access poly (car poly-lst) :parents) pt)))
(length-of-shortest-polys-with-var (cdr poly-lst) pt
(length (access poly
(car poly-lst)
:alist))))
(t
(length-of-shortest-polys-with-var (cdr poly-lst) pt n))))
(defun shortest-polys-with-var1 (poly-lst pt n)
(cond ((or (null poly-lst)
(eq n t))
nil)
((and (or (equal (length (access poly (car poly-lst) :alist)) n)
(equal (length (access poly (car poly-lst) :alist)) (+ n 1)))
(not (ignore-polyp (access poly (car poly-lst) :parents) pt)))
(cons (car poly-lst)
(shortest-polys-with-var1 (cdr poly-lst) pt n)))
(t
(shortest-polys-with-var1 (cdr poly-lst) pt n))))
(defun shortest-polys-with-var (var pot-lst pt)
; Var is a pot-label in pot-lst and pt is a parent tree of polys to ignore.
; We return a list of the polys with the shortest alists in the pot
; labeled with var. These polys can be considered as the ``simplest''
; polys about var.
(cond ((null pot-lst)
nil)
((equal var (access linear-pot (car pot-lst) :var))
; We have found the pot we are looking for. We find the length of the
; shortest polys in the pot. We then find all the polys in the pot
; with alists of that length, and return a list of those polys.
(let ((n (length-of-shortest-polys-with-var
(append (access linear-pot
(car pot-lst)
:negatives)
(access linear-pot
(car pot-lst)
:positives))
pt
t)))
(append (shortest-polys-with-var1
(access linear-pot (car pot-lst) :negatives)
pt n)
(shortest-polys-with-var1
(access linear-pot (car pot-lst) :positives)
pt n))))
(t (shortest-polys-with-var var (cdr pot-lst) pt))))
(defun binary-*-leaves (term)
(if (eq (fn-symb term) 'BINARY-*)
(cons (fargn term 1)
(binary-*-leaves (fargn term 2)))
(list term)))
(defun binary-*-tree (leaves)
; Return the BINARY-* term with leaves at its tips. In practice,
; leaves always contains at least two elements.
(cond ((null (cdr leaves))
(car leaves))
((null (cddr leaves))
(cons-term 'BINARY-*
(list
(car leaves)
(cadr leaves))))
(t
(cons-term 'BINARY-*
(list
(car leaves)
(binary-*-tree (cdr leaves)))))))
(defun merge-arith-term-order (l1 l2)
(cond ((null l1) l2)
((null l2) l1)
((arith-term-order (car l2) (car l1))
(cons (car l2) (merge-arith-term-order l1 (cdr l2))))
(t (cons (car l1) (merge-arith-term-order (cdr l1) l2)))))
(defun insert-arith-term-order (item list)
(cond ((null list)
(list item))
((arith-term-order item (car list))
(cons item list))
(t
(cons (car list)
(insert-arith-term-order item (cdr list))))))
(defun sort-arith-term-order (list)
(cond ((null list)
nil)
(t
(insert-arith-term-order (car list)
(sort-arith-term-order (cdr list))))))
(defun multiply-alist-and-const (alist const poly)
(cond ((null alist)
poly)
(t
(let ((temp-poly (add-linear-term (cons-term 'BINARY-*
(list
(kwote (* const
(cdar alist)))
(caar alist)))
'rhs
poly)))
(multiply-alist-and-const (cdr alist)
const
temp-poly)))))
(defun collect-rational-poly-p-lst (poly-lst)
(cond ((endp poly-lst) nil)
((access poly (car poly-lst) :rational-poly-p)
(cons (car poly-lst)
(collect-rational-poly-p-lst (cdr poly-lst))))
(t (collect-rational-poly-p-lst (cdr poly-lst)))))