-
Notifications
You must be signed in to change notification settings - Fork 0
/
Copy pathchap-filt-order.tex
1141 lines (1034 loc) · 54.7 KB
/
chap-filt-order.tex
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
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
\chapter{Orderings of filters in terms of reloids}
Whilst the other chapters of this book use filters to research funcoids
and reloids, here the opposite thing is discussed, the theory of reloids
is used to describe properties of filters.
In this chapter the word \emph{filter} is used to denote a filter
on a set (not on an arbitrary poset) only.
\section{Ordering of filters}
Below I will define some categories having filters (with possibly
different bases) as their objects and some relations having two filters
(with possibly different bases) as arguments induced by these categories
(defined as existence of a morphism between these two filters).
\begin{thm}
$\card a=\card U$ for every ultrafilter $a$ on $U$ if $U$ is infinite.\end{thm}
\begin{proof}
Let $f(X)=X$ if $X\in a$ and $f(X)=U\setminus X$ if $X\notin a$.
Obviously $f$ is a surjection from $U$ to $a$.
Every $X\in a$ appears as a value of $f$ exactly twice, as $f(X)$
and $f(U\setminus X)$. So $\card a=(\card U)/2=\card U$.\end{proof}
\begin{cor}
Cardinality of every two ultrafilters on a set $U$ is the same.\end{cor}
\begin{proof}
For infinite $U$ it follows from the theorem. For finite case it
is obvious.\end{proof}
\begin{prop}
$\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}=\setcond{C\in\subsets(\Dst f)}{\rsupfun{f^{-1}}C\in\mathcal{A}}$
for every $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$.
(Here a funcoid is considered as a pair of functions $\mathfrak{F}(\Base(\mathcal{A}))\rightarrow\mathfrak{F}(\Base(\mathcal{B}))$,
$\mathfrak{F}(\Base(\mathcal{B}))\rightarrow\mathfrak{F}(\Base(\mathcal{A}))$
rather than as a pair of functions $\mathscr{F}(\Base(\mathcal{A}))\rightarrow\mathscr{F}(\Base(\mathcal{B}))$,
$\mathscr{F}(\Base(\mathcal{B}))\rightarrow\mathscr{F}(\Base(\mathcal{A}))$.)\end{prop}
\begin{proof}
For every set $C\in\subsets\Base(\mathcal{B})$ we have
\begin{align*}
\rsupfun{f^{-1}}C\in\mathcal{A} & \Rightarrow\\
\exists K\in\mathcal{A}:\rsupfun{f^{-1}}C=K & \Rightarrow\\
\exists K\in\mathcal{A}:\rsupfun f\rsupfun{f^{-1}}C=\rsupfun fK & \Rightarrow\\
\exists K\in\mathcal{A}:C\supseteq\rsupfun fK & \Leftrightarrow\\
\exists K\in\mathcal{A}:C\in\rsupfun{\uparrow^{\mathsf{FCD}}f}K & \Rightarrow\\
C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}.
\end{align*}
So $C\in \setcond{C\in\subsets(\Dst f)}{\rsupfun{f^{-1}}C\in\mathcal{A}}\Rightarrow C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
Let now $C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$. Then
$\uparrow\rsupfun{f^{-1}}C\sqsupseteq\supfun{\uparrow^{\mathsf{FCD}}f^{-1}}\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}\sqsupseteq\mathcal{A}$
and thus $\rsupfun{f^{-1}}C\in\mathcal{A}$.
\end{proof}
Below I'll define some directed multigraphs. By an abuse of notation,
I will denote these multigraphs the same as (below defined) categories
based on some of these directed multigraphs with added composition
of morphisms (of directed multigraphs edges). As such I will call
vertices of these multigraphs objects and edges morphisms.
\begin{defn}
I will denote $\mathbf{GreFunc}{}_{1}$ the multigraph whose objects
are filters and whose morphisms between objects $\mathcal{A}$ and
$\mathcal{B}$ are $\mathbf{Set}$-morphisms from $\Base(\mathcal{A})$
to $\Base(\mathcal{B})$ such that $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
\end{defn}
\begin{defn}
I will denote $\mathbf{GreFunc}{}_{2}$ the multigraph whose objects
are filters and whose morphisms between objects $\mathcal{A}$ and
$\mathcal{B}$ are $\mathbf{Set}$-morphisms from $\Base(\mathcal{A})$
to $\Base(\mathcal{B})$ such that $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
\end{defn}
\begin{defn}
Let $\mathcal{A}$ be a filter on a set $X$ and $\mathcal{B}$ be
a filter on a set $Y$. $\mathcal{A}\ge_{1}\mathcal{B}$ iff $\Hom_{\mathbf{GreFunc}_{1}}(\mathcal{A},\mathcal{B})$
is not empty.
\end{defn}
\begin{defn}
Let $\mathcal{A}$ be a filter on a set $X$ and $\mathcal{B}$ be
a filter on a set $Y$. $\mathcal{A}\ge_{2}\mathcal{B}$ iff $\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})$
is not empty.\end{defn}
\begin{prop}
~
\begin{enumerate}
\item \label{gre-imp}$f\in\Hom_{\mathbf{GreFunc}_{1}}(\mathcal{A},\mathcal{B})$
iff $f$ is a $\mathbf{Set}$-morphism from $\Base(\mathcal{A})$
to $\Base(\mathcal{B})$ such that
\[
C\in\mathcal{B}\Leftarrow\rsupfun{f^{-1}}C\in\mathcal{A}
\]
for every $C\in\subsets\Base(\mathcal{B})$.
\item \label{gre-eq}$f\in\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})$
iff $f$ is a $\mathbf{Set}$-morphism from $\Base(\mathcal{A})$
to $\Base(\mathcal{B})$ such that
\[
C\in\mathcal{B}\Leftrightarrow\rsupfun{f^{-1}}C\in\mathcal{A}
\]
for every $C\in\subsets\Base(\mathcal{B})$.
\end{enumerate}
\end{prop}
\begin{proof}
~
\begin{widedisorder}
\item [{\ref{gre-imp}}] ~
\begin{multline*}
f\in\Hom_{\mathbf{GreFunc}_{1}}(\mathcal{A},\mathcal{B})\Leftrightarrow\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}\Leftrightarrow\\
\forall C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}:C\in\mathcal{B}\Leftrightarrow\forall C\in\subsets\Base(\mathcal{B}):(\rsupfun{f^{-1}}C\in\mathcal{A}\Rightarrow C\in\mathcal{B}).
\end{multline*}
\item [{\ref{gre-eq}}] ~
\begin{multline*}
f\in\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})\Leftrightarrow\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}\Leftrightarrow\forall C:(C\in\mathcal{B}\Leftrightarrow C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A})\Leftrightarrow\\
\forall C\in\subsets\Base(\mathcal{B}):(C\in\mathcal{B}\Leftrightarrow C\in\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A})\Leftrightarrow\\
\forall C\in\subsets\Base(\mathcal{B}):(\rsupfun{f^{-1}}C\in\mathcal{A}\Leftrightarrow C\in\mathcal{B}).
\end{multline*}
\end{widedisorder}
\end{proof}
\begin{defn}
The directed multigraph $\mathbf{FuncBij}$ is the directed multigraph
got from $\mathbf{GreFunc}_{2}$ by restricting to only bijective
morphisms.
\end{defn}
\begin{defn}
\index{directly isomorphic}A filter $\mathcal{A}$ is \emph{directly
isomorphic} to a filter $\mathcal{B}$ iff there is a morphism $f\in\Hom_{\mathbf{FuncBij}}(\mathcal{A},\mathcal{B})$.\end{defn}
\begin{obvious}
$f\in\Hom_{\mathbf{GreFunc}_{1}}(\mathcal{A},\mathcal{B})\Leftrightarrow\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
for every $\mathbf{Set}$-morphism from $\Base(\mathcal{A})$ to $\Base(\mathcal{B})$.
\end{obvious}
\begin{obvious}
$f\in\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})\Leftrightarrow\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
for every $\mathbf{Set}$-morphism from $\Base(\mathcal{A})$ to $\Base(\mathcal{B})$.
\end{obvious}
\begin{cor}
$\mathcal{A}\ge_{1}\mathcal{B}$ iff it exists a $\mathbf{Set}$-morphism
$f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$ such that $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
\end{cor}
\begin{cor}
$\mathcal{A}\ge_{2}\mathcal{B}$ iff it exists a $\mathbf{Set}$-morphism
$f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$ such that $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.\end{cor}
\begin{prop}
For a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
the following are equivalent:
\begin{enumerate}
\item \label{fbij-star}$\mathcal{B}=\setcond{C\in\subsets\Base(\mathcal{B})}{\rsupfun{f^{-1}}C\in\mathcal{A}}$.
\item \label{fbij-eback}$\forall C\in\Base(\mathcal{B}):(C\in\mathcal{B}\Leftrightarrow\rsupfun{f^{-1}}C\in\mathcal{A})$.
\item \label{fbij-eforw}$\forall C\in\Base(\mathcal{A}):(C\in\rsupfun f\mathcal{B}\Leftrightarrow C\in\mathcal{A})$.
\item \label{fbij-rbij}$\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$
is a bijection from $\mathcal{A}$ to~$\mathcal{B}$.
\item \label{fbij-rsurj}$\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$
is a function onto~$\mathcal{B}$.
\item \label{fbij-feq}$\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
\item \label{fbij-gre}$f\in\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})$.
\item \label{fbij-grp}$f\in\Hom_{\mathbf{FuncBij}}(\mathcal{A},\mathcal{B})$.
\end{enumerate}
\end{prop}
\begin{proof}
~
\begin{description}
\item [{\ref{fbij-star}$\Leftrightarrow$\ref{fbij-eback}}] ~
\[
\mathcal{B}=\setcond{C\in\subsets\Base(\mathcal{B})}{\rsupfun{f^{-1}}C\in\mathcal{A}}\Leftrightarrow\\
\forall C\in\subsets\Base(\mathcal{B}):(C\in\mathcal{B}\Leftrightarrow\rsupfun{f^{-1}}C\in\mathcal{A}).
\]
\item [{\ref{fbij-eback}$\Leftrightarrow$\ref{fbij-eforw}}] Because
$f$ is a bijection.
\item [{\ref{fbij-eback}$\Rightarrow$\ref{fbij-rsurj}}] For every $C\in\mathcal{B}$
we have $\rsupfun{f^{-1}}C\in\mathcal{A}$ and thus $\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}\supfun{\uparrow^{\mathsf{FCD}}f^{-1}}C=\rsupfun f\rsupfun{f^{-1}}C=C$.
Thus $\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$ is onto $\mathcal{B}$.
\item [{\ref{fbij-rbij}$\Rightarrow$\ref{fbij-rsurj}}] Obvious.
\item [{\ref{fbij-rsurj}$\Rightarrow$\ref{fbij-rbij}}] We need to prove
only that $\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$ is an
injection. But this follows from the fact that $f$ is a bijection.
\item [{\ref{fbij-rbij}$\Rightarrow$\ref{fbij-eforw}}] We have $\forall C\in\Base(\mathcal{A}):((\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}})C\in\mathcal{B}\Leftrightarrow C\in\mathcal{A})$
and consequently $\forall C\in\Base(\mathcal{A}):(\rsupfun fC\in\mathcal{B}\Leftrightarrow C\in\mathcal{A})$.
\item [{\ref{fbij-feq}$\Leftrightarrow$\ref{fbij-star}}] From the last
corollary.
\item [{\ref{fbij-star}$\Leftrightarrow$\ref{fbij-gre}}] Obvious.
\item [{\ref{fbij-gre}$\Leftrightarrow$\ref{fbij-grp}}] Obvious.
\end{description}
\end{proof}
\begin{cor}
The following are equivalent for every filters $\mathcal{A}$ and
$\mathcal{B}$:
\begin{enumerate}
\item $\mathcal{A}$ is directly isomorphic to $\mathcal{B}$.
\item There is a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that for every $C\in\mathscr{P}\Base(\mathcal{B})$
\[
C\in\mathcal{B}\Leftrightarrow\rsupfun{f^{-1}}C\in\mathcal{A}.
\]
\item There is a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that for every $C\in\mathscr{P}\Base(\mathcal{B})$
\[
\rsupfun fC\in\mathcal{B}\Leftrightarrow C\in\mathcal{A}.
\]
\item There is a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$ is a
bijection from $\mathcal{A}$ to $\mathcal{B}$.
\item There is a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\supfun{\uparrow^{\mathsf{FCD}}f}|_{\mathcal{A}}$ is a
function onto $\mathcal{B}$.
\item There is a bijective $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
\item There is a bijective morphism $f\in\Hom_{\mathbf{GreFunc}_{2}}(\mathcal{A},\mathcal{B})$.
\item There is a bijective morphism $f\in\Hom_{\mathbf{FuncBij}}(\mathcal{A},\mathcal{B})$.
\end{enumerate}
\end{cor}
\begin{prop}
$\mathbf{GreFunc}_{1}$ and $\mathbf{GreFunc}_{2}$ with function
composition are categories.\end{prop}
\begin{proof}
Let $f:\mathcal{A}\rightarrow\mathcal{B}$ and $g:\mathcal{B}\rightarrow\mathcal{C}$
be morphisms of $\mathbf{GreFunc}_{1}$. Then $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
and $\mathcal{C}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}g}\mathcal{B}$.
So
\[
\supfun{\uparrow^{\mathsf{FCD}}(g\circ f)}\mathcal{A}=\supfun{\uparrow^{\mathsf{FCD}}g}\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}\sqsupseteq\supfun{\uparrow^{\mathsf{FCD}}g}\mathcal{B}\sqsupseteq\mathcal{C}.
\]
Thus $g\circ f$ is a morphism of $\mathbf{GreFunc}_{1}$. Associativity
law is evident. $\id_{\Base(\mathcal{A})}$ is the identity morphism
of $\mathbf{GreFunc}_{1}$ for every filter~$\mathcal{A}$.
Let $f:\mathcal{A}\rightarrow\mathcal{B}$ and $g:\mathcal{B}\rightarrow\mathcal{C}$
be morphisms of $\mathbf{GreFunc}_{2}$. Then $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
and $\mathcal{C}=\supfun{\uparrow^{\mathsf{FCD}}g}\mathcal{B}$. So
\[
\supfun{\uparrow^{\mathsf{FCD}}(g\circ f)}\mathcal{A}=\supfun{\uparrow^{\mathsf{FCD}}g}\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}=\supfun{\uparrow^{\mathsf{FCD}}g}\mathcal{B}=\mathcal{C}.
\]
Thus $g\circ f$ is a morphism of $\mathbf{GreFunc}_{2}$. Associativity
law is evident. $\id_{\Base(\mathcal{A})}$ is the identity morphism
of $\mathbf{GreFunc}_{2}$ for every filter~$\mathcal{A}$.\end{proof}
\begin{cor}
$\le_{1}$ and $\le_{2}$ are preorders. \end{cor}
\begin{thm}
$\mathbf{FuncBij}$ is a groupoid.\end{thm}
\begin{proof}
First let's prove it is a category. Let $f:\mathcal{A}\rightarrow\mathcal{B}$
and $g:\mathcal{B}\rightarrow\mathcal{C}$ be morphisms of $\mathbf{FuncBij}$.
Then $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$ and $g:\Base(\mathcal{B})\rightarrow\Base(\mathcal{C})$
are bijections and $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
and $\mathcal{C}=\supfun{\uparrow^{\mathsf{FCD}}g}\mathcal{B}$. Thus
$g\circ f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{C})$ is a bijection
and $\mathcal{C}=\supfun{\uparrow^{\mathsf{FCD}}(g\circ f)}\mathcal{A}$.
Thus $g\circ f$ is a morphism of $\mathbf{FuncBij}$. $\id_{\Base(\mathcal{A})}$
is the identity morphism of $\mathbf{FuncBij}$ for every filter $\mathcal{A}$.
Thus it is a category.
It remains to prove only that every morphism $f\in\Hom_{\mathbf{FuncBij}}(\mathcal{A},\mathcal{B})$
has a reverse (for every filters $\mathcal{A}$, $\mathcal{B}$).
We have $f$ is a bijection $\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that for every $C\in\subsets\Base(\mathcal{A})$
\[
\rsupfun fC\in\mathcal{B}\Leftrightarrow C\in\mathcal{A}.
\]
Then $f^{-1}:\Base(\mathcal{B})\rightarrow\Base(\mathcal{A})$ is
a bijection such that for every $C\in\subsets\Base(\mathcal{B})$
\[
\rsupfun{f^{-1}}C\in\mathcal{A}\Leftrightarrow C\in\mathcal{B}.
\]
Thus $f^{-1}\in\Hom_{\mathbf{FuncBij}}(\mathcal{B},\mathcal{A})$.\end{proof}
\begin{cor}
Being directly isomorphic is an equivalence relation.
\end{cor}
\index{order!Rudin-Keisler}Rudin-Keisler order of ultrafilters is
considered in such a book as \cite{comfort-ultra}.
\begin{obvious}
For the case of ultrafilters being directly isomorphic is the same
as being Rudin-Keisler equivalent.\end{obvious}
\begin{defn}
\index{isomorphic!filters}A filter $\mathcal{A}$ is \emph{isomorphic}
to a filter $\mathcal{B}$ iff there exist sets $A\in\mathcal{A}$
and $B\in\mathcal{B}$ such that $\mathcal{A}\div A$ is directly
isomorphic to $\mathcal{B}\div B$.\end{defn}
\begin{obvious}
Equivalent filters are isomorphic.\end{obvious}
\begin{thm}
Being isomorphic (for small filters) is an equivalence relation.\end{thm}
\begin{proof}
~
\begin{description}
\item [{Reflexivity}] Because every filter is directly isomorphic to itself.
\item [{Symmetry}] If filter $\mathcal{A}$ is isomorphic to $\mathcal{B}$
then there exist sets $A\in\mathcal{A}$ and $B\in\mathcal{B}$ such
that $\mathcal{A}\div A$ is directly isomorphic to $\mathcal{B}\div B$
and thus $\mathcal{B}\div B$ is directly isomorphic to $\mathcal{A}\div A$.
So $\mathcal{B}$ is isomorphic to $\mathcal{A}$.
\item [{Transitivity}] Let $\mathcal{A}$ be isomorphic to $\mathcal{B}$
and $\mathcal{B}$ be isomorphic to $\mathcal{C}$. Then exist $A\in\mathcal{A}$,
$B_{1}\in\mathcal{B}$, $B_{2}\in\mathcal{B}$, $C\in\mathcal{C}$
such that there are bijections $f:A\rightarrow B_{1}$ and $g:B_{2}\rightarrow C$
such that
\[
\forall X\in\subsets A:(X\in\mathcal{B}\Leftrightarrow\rsupfun{f^{-1}}X\in\mathcal{A})\quad\text{and}\quad\forall X\in\subsets B_{1}:(X\in\mathcal{A}\Leftrightarrow\rsupfun fX\in\mathcal{B})
\]
and also $\forall X\in\subsets B_{2}:(X\in\mathcal{B}\Leftrightarrow\rsupfun gX\in\mathcal{C})$.
So $g\circ f$ is a bijection from $\rsupfun{f^{-1}}(B_{1}\cap B_{2})\in\mathcal{A}$
to $\rsupfun g(B_{1}\cap B_{2})\in\mathcal{C}$ such that
\[
X\in\mathcal{A}\Leftrightarrow\rsupfun fX\in\mathcal{B}\Leftrightarrow\rsupfun g\rsupfun fX\in\mathcal{C}\Leftrightarrow\rsupfun{g\circ f}X\in\mathcal{C}.
\]
Thus $g\circ f$ establishes a bijection which proves that $\mathcal{A}$
is isomorphic to~$\mathcal{C}$.
\end{description}
\end{proof}
\begin{lem}
Let $\card X=\card Y$, $u$ be an ultrafilter on $X$ and $v$ be
an ultrafilter on $Y$; let $A\in u$ and $B\in v$. Let $u\div A$
and $v\div B$ be directly isomorphic. Then if $\card(X\setminus A)=\card(Y\setminus B)$
we have $u$ and $v$ directly isomorphic.\end{lem}
\begin{proof}
Arbitrary extend the bijection witnessing being directly isomorphic
to the sets $X\setminus A$ and $X\setminus B$.\end{proof}
\begin{thm}
If $\card X=\card Y$ then being isomorphic and being directly isomorphic
are the same for ultrafilters $u$ on $X$ and $v$ on $Y$.\end{thm}
\begin{proof}
That if two filters are isomorphic then they are directly isomorphic
is obvious.
Let ultrafilters $u$ and $v$ be isomorphic that is there is a bijection
$f:A\rightarrow B$ where $A\in u$, $B\in v$ witnessing isomorphism
of $u$ and $v$.
If one of the filters $u$ or $v$ is a trivial ultrafilter then the
other is also a trivial ultrafilter and as it is easy to show they
are directly isomorphic. So we can assume $u$ and $v$ are not trivial
ultrafilters.
If $\card(X\setminus A)=\card(Y\setminus B)$ our statement follows
from the last lemma.
Now assume without loss of generality $\card(X\setminus A)<\card(Y\setminus B)$.
$\card B=\card Y$ because otherwise $\card(X\setminus A)=\card(Y\setminus B)$.
It is easy to show that there exists $B'\supset B$ such that $\card(X\setminus A)=\card(Y\setminus B')$
and $\card B'=\card B$.
We will find a bijection $g$ from $B$ to $B'$ which witnesses direct
isomorphism of $v$ to $v$ itself. Then the composition $g\circ f$
witnesses a direct isomorphism of $u\div A$ and $v\div B'$ and by
the lemma $u$ and $v$ are directly isomorphic.
Let $D=B'\setminus B$. We have $D\notin v$.
There exists a set $E\subseteq B$ such that $\card E\ge\card D$
and $E\notin v$.
We have $\card E=\card(D\cup E)$ and thus there exists a bijection
$h:E\rightarrow D\cup E$.
Let
\[
g(x)=\begin{cases}
x & \text{if }x\in B\setminus E;\\
h(x) & \text{if }x\in E.
\end{cases}
\]
$g|_{B\setminus E}$ and $g|_{E}$ are bijections.
$\im(g|_{B\setminus E})=B\setminus E$; $\im(g|_{E})=\im h=D\cup E$;
\[
(D\cup E)\cap(B\setminus E)=(D\cap(B\setminus E))\cup(E\cap(B\setminus E))=\emptyset\cup\emptyset=\emptyset.
\]
Thus $g$ is a bijection from $B$ to $(B\setminus E)\cup(D\cup E)=B\cup D=B'$.
To finish the proof it's enough to show that $\rsupfun gv=v$. Indeed
it follows from $B\setminus E\in v$.\end{proof}
\begin{prop}
~
\begin{enumerate}
\item \label{ge2-restr}For every $A\in\mathcal{A}$ and $B\in\mathcal{B}$
we have $\mathcal{A}\ge_{2}\mathcal{B}$ iff $\mathcal{A}\div A\ge_{2}\mathcal{B}\div B$.
\item \label{ge1-restr}For every $A\in\mathcal{A}$ and $B\in\mathcal{B}$
we have $\mathcal{A}\ge_{1}\mathcal{B}$ iff $\mathcal{A}\div A\ge_{1}\mathcal{B}\div B$.
\end{enumerate}
\end{prop}
\begin{proof}
~
\begin{widedisorder}
\item [{\ref{ge2-restr}}] $\mathcal{A}\ge_{2}\mathcal{B}$ iff there exist
a bijective $\mathbf{Set}$-morphism $f$ such that $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
The equality is obviously preserved replacing $\mathcal{A}$ with
$\mathcal{A}\div A$ and $\mathcal{B}$ with $\mathcal{B}\div B$.
\item [{\ref{ge1-restr}}] $\mathcal{A}\ge_{1}\mathcal{B}$ iff there exist
a bijective $\mathbf{Set}$-morphism $f$ such that $\mathcal{B}\subseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
The equality is obviously preserved replacing $\mathcal{A}$ with
$\mathcal{A}\div A$ and $\mathcal{B}$ with $\mathcal{B}\div B$.
\end{widedisorder}
\end{proof}
\begin{prop}
For ultrafilters $\ge_{2}$ is the same as Rudin-Keisler ordering
(as defined in \cite{comfort-ultra}).\end{prop}
\begin{proof}
$x\ge_{2}y$ iff there exist sets $A\in x$ and $B\in y$ and a bijective
$\mathbf{Set}$-morphism $f:X\rightarrow Y$ such that
\[
y\div B=\setcond{C\in\subsets Y}{\rsupfun{f^{-1}}C\in x\div A}
\]
that is when $C\in y\div B\Leftrightarrow\rsupfun{f^{-1}}C\in x\div A$
what is equivalent to~$C\in y\Leftrightarrow\rsupfun{f^{-1}}C\in x$
what is the definition of Rudin-Keisler ordering.\end{proof}
\begin{rem}
\index{Rudin-Keisler equivalence}The relation of being isomorphic
for ultrafilters is traditionally called \emph{Rudin-Keisler equivalence}.\end{rem}
\begin{obvious}
$(\ge_{1})\supseteq(\ge_{2})$.\end{obvious}
\begin{defn}
Let $Q$ and $R$ be binary relations on the set of (small) filters. I will
denote $\mathbf{MonRld}_{Q,R}$ the directed multigraph with objects
being filters and morphisms such monovalued reloids $f$ that $(\dom f)\mathrel Q\mathcal{A}$
and $(\im f)\mathrel R\mathcal{B}$.
I will also denote $\mathbf{CoMonRld}_{Q,R}$ the directed multigraph
with objects being filters and morphisms such injective reloids $f$
that $(\im f)\mathrel Q\mathcal{A}$ and $(\dom f)\mathrel R\mathcal{B}$.
These are essentially the duals.
\end{defn}
Some of these directed multigraphs are categories with reloid composition
(see below). By abuse of notation I will denote these categories the
same as these directed multigraphs.
\begin{lem}
$\mathbf{CoMonRld}_{Q,R}\ne\emptyset \Leftrightarrow\mathbf{MonRld}_{Q,R}\ne\emptyset$.
\end{lem}
\begin{proof}
\begin{multline*}
f\in\mathbf{CoMonRld}_{Q,R} \Leftrightarrow
(\im f)\mathrel Q\mathcal{A} \land (\dom f)\mathrel R\mathcal{B} \Leftrightarrow \\
(\dom f^{-1})\mathrel Q\mathcal{A} \land (\im f^{-1})\mathrel R\mathcal{B} \Leftrightarrow
f^{-1}\in\mathbf{MonRld}_{Q,R}
\end{multline*}
for every monovalued reloid~$f$ (or what is the same, injective reloid~$f^{-1}$).
\end{proof}
\begin{thm}
For every filters $\mathcal{A}$ and $\mathcal{B}$ the following
are equivalent:
\begin{enumerate}
\item \label{ge1-ineq}$\mathcal{A}\ge_{1}\mathcal{B}$.
\item \label{ge1-eq-ge}$\Hom_{\mathbf{MonRld}_{=,\sqsupseteq}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{ge1-le-ge}$\Hom_{\mathbf{MonRld}_{\sqsubseteq,\sqsupseteq}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{ge1-le-eq}$\Hom_{\mathbf{MonRld}_{\sqsubseteq,=}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{g1-c-eq-ge}$\Hom_{\mathbf{CoMonRld}_{=,\sqsupseteq}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{g1-c-le-ge}$\Hom_{\mathbf{CoMonRld}_{\sqsubseteq,\sqsupseteq}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{g1-c-le-eq}$\Hom_{\mathbf{CoMonRld}_{\sqsubseteq,=}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\end{enumerate}
\end{thm}
\begin{proof}
~
\begin{description}
\item [{\ref{ge1-ineq}$\Rightarrow$\ref{ge1-eq-ge}}] There exists a
$\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
We have
\[
\dom(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\mathcal{A}\sqcap\top(\Base(\mathcal{A}))=\mathcal{A}
\]
and
\[
\im(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\im\tofcd(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\im(\uparrow^{\mathsf{FCD}}f)|_{\mathcal{A}}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}\sqsupseteq\mathcal{B}.
\]
Thus $(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}$ is a monovalued
reloid such that $\dom(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\mathcal{A}$
and $\im(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}\sqsupseteq\mathcal{B}$.
\item [{\ref{ge1-eq-ge}$\Rightarrow$\ref{ge1-le-ge},~\ref{ge1-le-eq}$\Rightarrow$\ref{ge1-le-ge},~\ref{g1-c-eq-ge}$\Rightarrow$\ref{g1-c-le-ge},~\ref{g1-c-le-eq}$\Rightarrow$\ref{g1-c-le-ge}}] Obvious.
\item [{\ref{ge1-le-ge}$\Rightarrow$\ref{ge1-ineq}}] We have $\mathcal{B}\sqsubseteq\supfun{\tofcd f}\mathcal{A}$
for a monovalued reloid $f\in\mathsf{RLD}(\Base(\mathcal{A}),\Base(\mathcal{B}))$.
Then there exists a $\mathbf{Set}$-morphism $F:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}F}\mathcal{A}$
that is $\mathcal{A}\ge_{1}\mathcal{B}$.
\item [{\ref{g1-c-le-ge}$\Rightarrow$\ref{g1-c-le-eq}}] Let $f$ be an injective reloid such that
$\im f\sqsubseteq\mathcal{A}$ and $\dom f\sqsupseteq\mathcal{B}$. Then
$\im f|_{\mathcal{B}}\sqsubseteq\mathcal{A}$ and $\dom f|_{\mathcal{B}}=\mathcal{B}$.
So $f|_{\mathcal{B}}\in\Hom_{\mathbf{CoMonRld}_{\sqsubseteq,=}}(\mathcal{A},\mathcal{B})$.
\item [{\ref{ge1-eq-ge}$\Leftrightarrow$\ref{g1-c-eq-ge},~\ref{ge1-le-ge}$\Leftrightarrow$\ref{g1-c-le-ge},~\ref{ge1-le-eq}$\Leftrightarrow$\ref{g1-c-le-eq}}] By
the lemma.
\end{description}
\end{proof}
\begin{thm}
For every filters $\mathcal{A}$ and $\mathcal{B}$ the following
are equivalent:
\begin{enumerate}
\item \label{ge2-in}$\mathcal{A}\ge_{2}\mathcal{B}$.
\item \label{ge2-mon}$\Hom_{\mathbf{MonRld}_{=,=}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\item \label{ge2-comon}$\Hom_{\mathbf{CoMonRld}_{=,=}}(\mathcal{A},\mathcal{B})\ne\emptyset$.
\end{enumerate}
\end{thm}
\begin{proof}
~
\begin{description}
\item [{\ref{ge2-in}$\Rightarrow$\ref{ge2-mon}}] Let $\mathcal{A}\ge_{2}\mathcal{B}$
that is $\mathcal{B}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
for some $\mathbf{Set}$-morphism $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$.
Then $\dom(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\mathcal{A}$
and
\[
\im(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\im\tofcd(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}=\im(\uparrow^{\mathsf{FCD}}f)|_{\mathcal{A}}=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}=\mathcal{B}.
\]
So $(\uparrow^{\mathsf{RLD}}f)|_{\mathcal{A}}$ is a sought for reloid.
\item [{\ref{ge2-mon}$\Rightarrow$\ref{ge2-in}}] There exists a monovalued reloid~$f$ with domain~$\mathcal{A}$
such that $\supfun{\tofcd f}\mathcal{A}=\mathcal{B}$.
By corollary \ref{mv-is-restr}
below, there exists a $\mathbf{Set}$-morphism $F:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $f=(\uparrow^{\mathsf{RLD}}F)|_{\mathcal{A}}$. Thus
\[
\supfun{\uparrow^{\mathsf{FCD}}F}\mathcal{A}=\im(\uparrow^{\mathsf{FCD}}F)|_{\mathcal{A}}=\im\tofcd(\uparrow^{\mathsf{RLD}}F)|_{\mathcal{A}}=\im\tofcd f=\im f=\mathcal{B}.
\]
Thus $\mathcal{A}\ge_{2}\mathcal{B}$ is testified by the morphism
$F$.
\item [{\ref{ge2-mon}$\Leftrightarrow$\ref{ge2-comon}}] By the lemma.
\end{description}
\end{proof}
\begin{thm}
The following are categories (with reloid composition):
\begin{enumerate}
\item \label{monrld-le-ge}$\mathbf{MonRld}_{\sqsubseteq,\sqsupseteq}$;
\item \label{monrld-le-eq}$\mathbf{MonRld}_{\sqsubseteq,=}$;
\item \label{monrld-eq-eq}$\mathbf{MonRld}_{=,=}$;
\item $\mathbf{CoMonRld}_{\sqsubseteq,\sqsupseteq}$;
\item $\mathbf{CoMonRld}_{\sqsubseteq,=}$;
\item $\mathbf{CoMonRld}_{=,=}$.
\end{enumerate}
\end{thm}
\begin{proof}
We will prove only the first three. The rest follow from duality.
We need to prove only that composition of morphisms is a morphism,
because associativity and existence of identity morphism are evident.
We have:
\begin{widedisorder}
\item [{\ref{monrld-le-ge}}] Let $f\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,\sqsupseteq}}(\mathcal{A},\mathcal{B})$,
$g\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,\sqsupseteq}}(\mathcal{B},\mathcal{C})$.
Then $\dom f\sqsubseteq\mathcal{A}$, $\im f\sqsupseteq\mathcal{B}$,
$\dom g\sqsubseteq\mathcal{B}$, $\im g\sqsupseteq\mathcal{C}$. So
$\dom(g\circ f)\sqsubseteq\mathcal{A}$, $\im(g\circ f)\sqsupseteq\mathcal{C}$
that is $g\circ f\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,\sqsupseteq}}(\mathcal{A},\mathcal{C})$.
\item [{\ref{monrld-le-eq}}] Let $f\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,=}}(\mathcal{A},\mathcal{B})$,
$g\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,=}}(\mathcal{B},\mathcal{C})$.
Then $\dom f\sqsubseteq\mathcal{A}$, $\im f=\mathcal{B}$, $\dom g\sqsubseteq\mathcal{B}$,
$\im g=\mathcal{C}$. So $\dom(g\circ f)\sqsubseteq\mathcal{A}$,
$\im(g\circ f)=\mathcal{C}$ that is $g\circ f\in\Hom_{\mathbf{MonRld}_{\sqsubseteq,=}}(\mathcal{A},\mathcal{C})$.
\item [{\ref{monrld-eq-eq}}] Let $f\in\Hom_{\mathbf{MonRld}_{=,=}}(\mathcal{A},\mathcal{B})$,
$g\in\Hom_{\mathbf{MonRld}_{=,=}}(\mathcal{B},\mathcal{C})$. Then
$\dom f=\mathcal{A}$, $\im f=\mathcal{B}$, $\dom g=\mathcal{B}$,
$\im g=\mathcal{C}$. So $\dom(g\circ f)=\mathcal{A}$, $\im(g\circ f)=\mathcal{C}$
that is $g\circ f\in\Hom_{\mathbf{MonRld}_{=,=}}(\mathcal{A},\mathcal{C})$.
\end{widedisorder}
\end{proof}
\begin{defn}
Let $\mathbf{BijRld}$ be the groupoid of all bijections of the category
of reloid triples. Its objects are filters and its morphisms from
a filter $\mathcal{A}$ to filter $\mathcal{B}$ are monovalued injective
reloids $f$ such that $\dom f=\mathcal{A}$ and $\im f=\mathcal{B}$.\end{defn}
\begin{thm}
Filters $\mathcal{A}$ and $\mathcal{B}$ are isomorphic iff $\Hom_{\mathbf{BijRld}}(\mathcal{A},\mathcal{B})\neq\emptyset$.\end{thm}
\begin{proof}
~
\begin{description}
\item [{$\Rightarrow$}] Let $\mathcal{A}$ and $\mathcal{B}$ be isomorphic.
Then there are sets $A\in\mathcal{A}$, $B\in\mathcal{B}$ and a bijective
$\mathbf{Set}$-morphism $F:A\rightarrow B$ such that $\rsupfun F:\subsets A\cap\mathcal{A}\rightarrow\subsets B\cap\mathcal{B}$
is a bijection.
Obviously $f=(\uparrow^{\mathsf{RLD}}F)|_{\mathcal{A}}$ is monovalued
and injective.
\begin{align*}
\im f & =\\
\bigsqcap^{\mathfrak{F}}\setcond{\im G}{G\in\up(\uparrow^{\mathsf{RLD}}F)|_{\mathcal{A}}} & =\\
\bigsqcap^{\mathfrak{F}}\setcond{\im(H\cap F|_{X})}{H\in\up(\uparrow^{\mathsf{RLD}}F)|_{\mathcal{A}},X\in\mathcal{A}} & =\\
\bigsqcap^{\mathfrak{F}}\setcond{\im F|_{P}}{P\in\mathcal{A}} & =\\
\bigsqcap^{\mathfrak{F}}\setcond{\rsupfun FP}{P\in\mathcal{A}} & =\\
\bigsqcap^{\mathfrak{F}}\setcond{\rsupfun FP}{P\in\subsets A\cap\mathcal{A}} & =\\
\bigsqcap^{\mathfrak{F}}(\subsets B\cap\mathcal{B}) & =\\
\bigsqcap^{\mathfrak{F}}\mathcal{B}=\mathcal{B}.
\end{align*}
Thus $\dom f=\mathcal{A}$ and $\im f=\mathcal{B}$.
\item [{$\Leftarrow$}] Let $f$ be a monovalued injective reloid such
that $\dom f=\mathcal{A}$ and $\im f=\mathcal{B}$. Then there exist
a function $F'$ and an injective binary relation $F''$ such that
$F',F''\in f$. Thus $F=F'\cap F''$ is an injection such that $F\in f$.
The function $F$ is a bijection from $A=\dom F$ to $B=\im F$. The
function $\rsupfun F$ is an injection on $\subsets A\cap\mathcal{A}$
(and moreover on $\subsets A$). It's simple to show that $\forall X\in\subsets A\cap\mathcal{A}:\rsupfun FX\in\subsets B\cap\mathcal{B}$
and similarly
\[
\forall Y\in\subsets B\cap\mathcal{B}:(\rsupfun F)^{-1}Y=\rsupfun{F^{-1}}Y\in\subsets A\cap\mathcal{A}.
\]
Thus $\rsupfun F|_{\subsets A\cap\mathcal{A}}$ is a bijection $\subsets A\cap\mathcal{A}\rightarrow\subsets B\cap\mathcal{B}$.
So filters $\mathcal{A}$ and $\mathcal{B}$ are isomorphic.
\end{description}
\end{proof}
\begin{prop}
$(\ge_{1})=(\sqsupseteq)\circ(\ge_{2})$ (when we limit to small filters).\end{prop}
\begin{proof}
$\mathcal{A}\ge_{1}\mathcal{B}$ iff exists a function $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
But $\mathcal{B}\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$
is equivalent to $\exists\mathcal{B}'\in\mathscr{F}:(\mathcal{B}'\sqsupseteq\mathcal{B}\land\mathcal{B}'=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A})$.
So $\mathcal{A}\ge_{1}\mathcal{B}$ is equivalent to existence of
$\mathcal{B}'\in\mathscr{F}$ such that $\mathcal{B}'\sqsupseteq\mathcal{B}$
and existence of a function $f:\Base(\mathcal{A})\rightarrow\Base(\mathcal{B})$
such that $\mathcal{B}'=\supfun{\uparrow^{\mathsf{FCD}}f}\mathcal{A}$.
This is equivalent to $\mathcal{A}\mathrel{((\sqsupseteq)\circ(\ge_{2}))}\mathcal{B}$.\end{proof}
\begin{prop}
If $a$ and $b$ are ultrafilters then $b\ge_{1}a\Leftrightarrow b\ge_{2}a$.\end{prop}
\begin{proof}
We need to prove only $b\ge_{1}a\Rightarrow b\ge_{2}a$. If $b\ge_{1}a$
then there exists a monovalued reloid $f:\Base(b)\rightarrow\Base(a)$
such that $\dom f=b$ and $\im f\sqsupseteq a$. Then $\im f=\im\tofcd f\in\{\bot^{\mathscr{F}(\Base(a))}\}\cup\atoms^{\mathscr{F}(\Base(a))}$
because $\tofcd f$ is a monovalued funcoid. So $\im f=a$ (taken
into account $\im f\ne\bot^{\mathscr{F}(\Base(a))}$) and thus $b\ge_{2}a$\@.\end{proof}
\begin{cor}
For atomic filters $\ge_{1}$ is the same as $\ge_{2}$.
\end{cor}
Thus I will write simply $\ge$ for atomic filters.
\subsection{Existence of no more than one monovalued injective reloid for a given
pair of ultrafilters}
\subsubsection{The lemmas}
The lemmas in this section were provided to me by \noun{Robert Martin Solovay}
in \cite{solovay-on-identity}. They are based on \noun{Wistar Comfort}'s
work.
In this section we will assume $\mu$ is an ultrafilter on a set $I$
and function $f:I\rightarrow I$ has the property $X\in\mu\Leftrightarrow\rsupfun{f^{-1}}X\in\mu$.
\begin{lem}
\label{lem:one-reloid-first}If $X\in\mu$ then $X\cap\rsupfun fX\in\mu$.\end{lem}
\begin{proof}
If $\rsupfun fX\notin\mu$ then $X\subseteq\rsupfun{f^{-1}}\rsupfun fX\notin\mu$
and so $X\notin\mu$. Thus $X\in\mu\land\rsupfun fX\in\mu$ and consequently
$X\cap\rsupfun fX\in\mu$.
\end{proof}
We will say that $x$ is \emph{periodic} when $f^{n}(x)=x$ for some
positive integer $x$. The least such $n$ is called \emph{the period}
of $x$.
Let's define $x\sim y$ iff there exist $i,j\in\mathbb{N}$ such that
$f^{i}(x)=f^{j}(y)$. Trivially it is an equivalence relation. If
$x$ and $y$ are periodic, then $x\sim y$ iff exists $n\in\mathbb{N}$
such that $f^{n}(y)=x$.
Let $A=\setcond{x\in I}{x\text{ is periodic with period}>1}$.
We will show $A\notin\mu$. Let's assume $A\in\mu$.
Let a set $D\subseteq A$ contains (by the axiom of choice) exactly
one element from each equivalence class of $A$ defined by the relation
$\sim$.
Let $\alpha$ be a function $A\rightarrow\mathbb{N}$ defined as follows.
Let $x\in A$. Let $y$ be the unique element of $D$ such that $x\sim y$.
Let $\alpha(x)$ be the least $n\in\mathbb{N}$ such that $f^{n}(y)=x$.
Let $B_{0}=\setcond{x\in A}{\alpha(x)\text{ is even}}$ and $B_{1}=\setcond{x\in A}{\alpha(x)\text{ is odd}}$.
Let $B_{2}=\setcond{x\in A}{\alpha(x)=0}$.
\begin{lem}
$B_{0}\cap\rsupfun fB_{0}\subseteq B_{2}$.\end{lem}
\begin{proof}
If $x\in B_{0}\cap\rsupfun fB_{0}$ then for a minimal even $n$ and
$x=f(x')$ where $f^{m}(y')=x'$ for a minimal even $m$. Thus $f^{n}(y)=f(x')$
thus $y$ and $x'$ laying in the same equivalence class and thus
$y=y'$. So we have $f^{n}(y)=f^{m+1}(y)$. Thus $n\le m+1$ by minimality.
$x'$ lies on an orbit and thus $x'=f^{-1}(x)$ where by $f^{-1}$
I mean step backward on our orbit; $f^{m}(y)=f^{-1}(x)$ and thus
$x'=f^{n-1}(y)$ thus $n-1\ge m$ by minimality or $n=0$.
Thus $n=m+1$ what is impossible for even $n$ and $m$. We have a
contradiction what proves $B_{0}\cap\rsupfun fB_{0}\subseteq\emptyset$.
Remained the case $n=0$, then $x=f^{0}(y)$ and thus $\alpha(x)=0$.\end{proof}
\begin{lem}
$B_{1}\cap\rsupfun fB_{1}=\emptyset$.\end{lem}
\begin{proof}
Let $x\in B_{1}\cap\rsupfun fB_{1}$. Then $f^{n}(y)=x$ for an odd
$n$ and $x=f(x')$ where $f^{m}(y')=x'$ for an odd $m$. Thus $f^{n}(y)=f(x')$
thus $y$ and $x'$ laying in the same equivalence class and thus
$y=y'$. So we have $f^{n}(y)=f^{m+1}(y)$. Thus $n\le m+1$ by minimality.
$x'$ lies on an orbit and thus $x'=f^{-1}(x)$ where by $f^{-1}$
I mean step backward on our orbit;
$f^{m}(y)=f^{-1}(x)$ and thus $x'=f^{n-1}(y)$ thus $n-1\ge m$ by
minimality ($n=0$ is impossible because $n$ is odd).
Thus $n=m+1$ what is impossible for odd $n$ and $m$. We have a
contradiction what proves $B_{1}\cap\rsupfun fB_{1}=\emptyset$.\end{proof}
\begin{lem}
$B_{2}\cap\rsupfun fB_{2}=\emptyset$.\end{lem}
\begin{proof}
Let $x\in B_{2}\cap\rsupfun fB_{2}$. Then $x=y$ and $x'=y$ where
$x=f(x')$. Thus $x=f(x)$ and so $x\notin A$ what is impossible.\end{proof}
\begin{lem}
$A\notin\mu$.\end{lem}
\begin{proof}
Suppose $A\in\mu$.
Since $A\in\mu$ we have $B_{0}\in\mu$ or $B_{1}\in\mu$.
So either $B_{0}\cap\rsupfun fB_{0}\subseteq B_{2}$ or $B_{1}\cap\rsupfun fB_{1}\subseteq B_{2}$.
As such by the lemma \ref{lem:one-reloid-first} we have $B_{2}\in\mu$.
This is incompatible with $B_{2}\cap\rsupfun fB_{2}=\emptyset$. So
we got a contradiction.
\end{proof}
Let $C$ be the set of points $x$ which are not periodic but $f^{n}(x)$
is periodic for some positive $n$.
\begin{lem}
$C\notin\mu$.\end{lem}
\begin{proof}
Let $\beta$ be a function $C\rightarrow\mathbb{N}$ such that $\beta(x)$
is the least $n\in\mathbb{N}$ such that $f^{n}(x)$ is periodic.
Let $C_{0}=\setcond{x\in C}{\beta(x)\text{ is even}}$ and $C_{1}=\setcond{x\in C}{\beta(x)\text{ is odd}}$.
Obviously $C_{j}\cap\rsupfun fC_{j}=\emptyset$ for $j=0,1$. Hence
by lemma \ref{lem:one-reloid-first} we have $C_{0},C_{1}\notin\mu$
and thus $C=C_{0}\cup C_{1}\notin\mu$.
\end{proof}
Let $E$ be the set of $x\in I$ such that for no $n\in\mathbb{N}$
we have $f^{n}(x)$ periodic.
\begin{lem}
Let $x,y\in E$ be such that $f^{i}(x)=f^{j}(y)$ and $f^{i'}(x)=f^{j'}(y)$
for some $i,j,i',j'\in\mathbb{N}$. Then $i-j=i'-j'$.\end{lem}
\begin{proof}
$i\mapsto f^{i}(x)$ is a bijection.
So $y=f^{i-j}(y)$ and $y=f^{i'-j'}(y)$. Thus $f^{i-j}(y)=f^{i'-j'}(y)$
and so $i-j=i'-j'$.\end{proof}
\begin{lem}
$E\notin\mu$.\end{lem}
\begin{proof}
Let $D'\subseteq E$ be a subset of $E$ with exactly one element
from each equivalence class of the relation $\sim$ on $E$.
Define the function $\gamma:E\rightarrow\mathbb{Z}$ as follows. Let
$x\in E$. Let $y$ be the unique element of $D'$ such that $x\sim y$.
Choose $i,j\in\mathbb{N}$ such that $f^{i}(y)=f^{j}(x)$. Let $\gamma(x)=i-j$.
By the last lemma, $\gamma$ is well-defined.
It is clear that if $x\in E$ then $f(x)\in E$ and moreover $\gamma(f(x))=\gamma(x)+1$.
Let $E_{0}=\setcond{x\in E}{\gamma(x)\text{ is even}}$ and $E_{1}=\setcond{x\in E}{\gamma(x)\text{ is odd}}$.
We have $E_{0}\cap\rsupfun fE_{0}=\emptyset\notin\mu$ and hence $E_{0}\notin\mu$.
Similarly $E_{1}\notin\mu$.
Thus $E=E_{0}\cup E_{1}\notin\mu$.\end{proof}
\begin{lem}
$f$ is the identity function on a set in $\mu$.\end{lem}
\begin{proof}
We have shown $A,C,E\notin\mu$. But the points which lie in none
of these sets are exactly points periodic with period $1$ that is
fixed points of $f$. Thus the set of fixed points of $f$ belongs
to the filter $\mu$.
\end{proof}
\subsubsection{The main theorem and its consequences}
\begin{thm}
For every ultrafilter $a$ the morphism $(a,a,\id_{a}^{\mathsf{FCD}})$
is the only
\begin{enumerate}
\item \label{atom-oneiso}monovalued morphism of the category of reloid
triples from $a$ to $a$;
\item injective morphism of the category of reloid triples from $a$ to
$a$;
\item bijective morphism of the category of reloid triples from $a$ to
$a$.
\end{enumerate}
\end{thm}
\begin{proof}
We will prove only \ref{atom-oneiso} because the rest follow from
it.
Let $f$ be a monovalued morphism of reloid triples from~$a$ to~$a$.
Then it exists a $\mathbf{Set}$-morphism $F$ such that $F\in f$.
Trivially $\supfun{\uparrow^{\mathsf{FCD}}F}a\sqsupseteq a$ and thus
$\rsupfun FA\in a$ for every $A\in a$. Thus by the lemma we have
that $F$ is the identity function on a set in $a$ and so obviously
$f$ is an identity.\end{proof}
\begin{cor}
For every two atomic filters (with possibly different bases) $\mathcal{A}$
and $\mathcal{B}$ there exists at most one bijective reloid triple
from $\mathcal{A}$ to $\mathcal{B}$.\end{cor}
\begin{proof}
Suppose that $f$ and $g$ are two different bijective reloids from
$\mathcal{A}$ to $\mathcal{B}$. Then $g^{-1}\circ f$ is not the
identity reloid (otherwise $g^{-1}\circ f=\id_{\dom f}^{\mathsf{RLD}}$
and so $f=g$ because $f$ and~$g$ are isomorphisms). But $g^{-1}\circ f$ is a bijective reloid (as a composition
of bijective reloids) from $\mathcal{A}$ to $\mathcal{A}$ what is
impossible.
\end{proof}
\section{Rudin-Keisler equivalence and Rudin-Keisler order}
\begin{thm}
Atomic filters $a$ and $b$ (with possibly different bases) are isomorphic
iff $a\ge b\land b\ge a$.\end{thm}
\begin{proof}
Let $a\ge b\land b\ge a$. Then there are a monovalued reloids $f$
and $g$ such that $\dom f=a$ and $\im f=b$ and $\dom g=b$ and
$\im g=a$. Thus $g\circ f$ and $f\circ g$ are monovalued morphisms
from $a$ to $a$ and from $b$ to $b$. By the above we have $g\circ f=\id_{a}^{\mathsf{RLD}}$
and $f\circ g=\id_{b}^{\mathsf{RLD}}$ so $g=f^{-1}$ and $f^{-1}\circ f=\id_{a}^{\mathsf{RLD}}$
and $f\circ f^{-1}=\id_{b}^{\mathsf{RLD}}$. Thus $f$ is an injective
monovalued reloid from $a$ to $b$ and thus $a$ and $b$ are isomorphic.
\end{proof}
The last theorem cannot be generalized from atomic filters to arbitrary
filters, as it's shown by the following example:
\begin{example}
$\mathcal{A}\ge_{1}\mathcal{B}\wedge\mathcal{B}\ge_{1}\mathcal{A}$
but $\mathcal{A}$ is not isomorphic to $\mathcal{B}$ for some filters
$\mathcal{A}$ and $\mathcal{B}$.\end{example}
\begin{proof}
Consider $\mathcal{A}=\uparrow^{\mathbb{R}}[0;1]$ and $\mathcal{B}=\bigsqcap\setcond{\uparrow^{\mathbb{R}}[0;1+\epsilon[}{\epsilon>0}$.
Then the function $f=\mylambda x{\mathbb{R}}{x/2}$ witnesses both
inequalities $\mathcal{A}\ge_{1}\mathcal{B}$ and $\mathcal{B}\ge_{1}\mathcal{A}$.
But these filters cannot be isomorphic because only one of them is
principal.\end{proof}
\begin{lem}
Let $f_{0}$ and $f_{1}$ be $\mathbf{Set}$-morphisms. Let $f(x,y)=(f_{0}x,f_{1}y)$
for a function $f$. Then
\[
\supfun{\uparrow^{\mathsf{FCD}(\Src f_{0}\times\Src f_{1},\Dst f_{0}\times\Dst f_{1})}f}(\mathcal{A}\times^{\mathsf{RLD}}\mathcal{B})=\supfun{\uparrow^{\mathsf{FCD}}f_{0}}\mathcal{A}\times^{\mathsf{RLD}}\supfun{\uparrow^{\mathsf{FCD}}f_{1}}\mathcal{B}.
\]
\end{lem}
\begin{proof}
~
\begin{align*}
\supfun{\uparrow^{\mathsf{FCD}(\Src f_{0}\times\Src f_{1},\Dst f_{0}\times\Dst f_{1})}f}(\mathcal{A}\times^{\mathsf{RLD}}\mathcal{B}) & =\\
\supfun{\uparrow^{\mathsf{FCD}(\Src f_{0}\times\Src f_{1},\Dst f_{0}\times\Dst f_{1})}f}\bigsqcap\setcond{\uparrow^{\Src f_{0}\times\Src f_{1}}(A\times B)}{A\in\mathcal{A},B\in\mathcal{B}} & =\\
\bigsqcap\setcond{\uparrow^{\Dst f_{0}\times\Dst f_{1}}\rsupfun f(A\times B)}{A\in\mathcal{A},B\in\mathcal{B}} & =\\
\bigsqcap\setcond{\uparrow^{\Dst f_{0}\times\Dst f_{1}}(\rsupfun{f_{0}}A\times\rsupfun{f_{1}}B)}{A\in\mathcal{A},B\in\mathcal{B}} & =\\
\bigsqcap\setcond{\uparrow^{\Dst f_{0}}\rsupfun{f_{0}}A\times\uparrow^{\Dst f_{1}}\rsupfun{f_{1}}B)}{A\in\mathcal{A},B\in\mathcal{B}} & =\text{ (theorem \ref{meet-prod-fcd})}\\
\bigsqcap\setcond{\uparrow^{\Dst f_{0}}\rsupfun{f_{0}}A}{A\in\mathcal{A}}\times^{\mathsf{RLD}}\bigsqcap\setcond{\uparrow^{\Dst f_{1}}\rsupfun{f_{1}}B}{B\in\mathcal{B}} & =\\
\supfun{\uparrow^{\mathsf{FCD}}f_{0}}\mathcal{A}\times^{\mathsf{RLD}}\supfun{\uparrow^{\mathsf{FCD}}f_{1}}\mathcal{B}.
\end{align*}
\end{proof}
\begin{thm}
\label{inj-iso-dom}Let $f$ be a monovalued reloid. Then $\GR f$
is isomorphic to the filter $\dom f$.\end{thm}
\begin{proof}
Let $f$ be a monovalued reloid. There exists a function $F\in\GR f$.
Consider the bijective function $p=\mylambda x{\dom F}{(x,Fx)}$.
$\rsupfun p\dom F=F$ and consequently
\begin{align*}
\supfun p\dom f & =\\
\bigsqcap_{K\in\up f}^{\mathsf{RLD}}\rsupfun p\dom K & =\\
\bigsqcap_{K\in\up f}^{\mathsf{RLD}}\rsupfun p\dom(K\cap F) & =\\
\bigsqcap_{K\in\up f}^{\mathsf{RLD}}(K\cap F) & =\\
\bigsqcap_{K\in\up f}^{\mathsf{RLD}}K & =f.
\end{align*}
Thus $p$ witnesses that $f$ is isomorphic to the filter $\dom f$.\end{proof}
\begin{cor}
The graph of a monovalued reloid with atomic domain is atomic.
\end{cor}
\begin{cor}
$\id_{\mathcal{A}}^{\mathsf{RLD}}$ is isomorphic to $\mathcal{A}$
for every filter $\mathcal{A}$.\end{cor}
\begin{thm}
There are atomic filters incomparable by Rudin-Keisler order. (Elements~$a$
and~$b$ are \emph{incomparable} when $a\nsqsubseteq b\land b\nsqsubseteq a$.)\end{thm}
\begin{proof}
See \cite{Gryzlov1997151}.\end{proof}
\begin{thm}
$\ge_{1}$ and $\ge_{2}$ are different relations.\end{thm}
\begin{proof}
Consider $a$ is an arbitrary non-empty filter. Then $a\ge_{1}\bot^{\mathscr{F}(\Base(a))}$
but not $a\ge_{2}\bot^{\mathscr{F}(\Base(a))}$.\end{proof}
\begin{prop}
If $a\ge_{2}b$ where $a$ is an ultrafilter then $b$ is also an
ultrafilter.\end{prop}
\begin{proof}
$b=\supfun{\uparrow^{\mathsf{FCD}}f}a$ for some $f:\Base(a)\rightarrow\Base(b)$.
So $b$ is an ultrafilter since $f$ is monovalued.\end{proof}
\begin{cor}
If $a\ge_{1}b$ where $a$ is an ultrafilter then $b$ is also an
ultrafilter or $\bot^{\mathscr{F}(\Base(a))}$.\end{cor}
\begin{proof}
$b\sqsubseteq\supfun{\uparrow^{\mathsf{FCD}}f}a$ for some $f:\Base(a)\rightarrow\Base(b)$.
Therefore $b'=\supfun{\uparrow^{\mathsf{FCD}}f}a$ is an ultrafilter.
From this our statement follows.\end{proof}
\begin{prop}
Principal filters, generated by sets of the same cardinality, are
isomorphic.\end{prop}
\begin{proof}
Let $A$ and $B$ be sets of the same cardinality. Then there are
a bijection $f$ from $A$ to $B$. We have $\rsupfun fA=B$ and thus
$A$ and $B$ are isomorphic.\end{proof}
\begin{prop}
If a filter is isomorphic to a principal filter, then it is also a
principal filter induced by a set with the same cardinality.\end{prop}
\begin{proof}
Let $A$ be a principal filter and $B$ is a filter isomorphic to
$A$. Then there are sets $X\in A$ and $Y\in B$ such that there
are a bijection $f:X\rightarrow Y$ such that $\rsupfun fA=B$.
So $\min B$ exists and $\min B=\rsupfun f\min A$ and thus $B$ is
a principal filter (of the same cardinality as $A$).\end{proof}
\begin{prop}
A filter isomorphic to a non-trivial ultrafilter is a non-trivial
ultrafilter.\end{prop}
\begin{proof}
Let $a$ be a non-trivial ultrafilter and $a$ be isomorphic to $b$.
Then $a\ge_{2}b$ and thus $b$ is an ultrafilter. The filter $b$
cannot be trivial because otherwise $a$ would be also trivial.\end{proof}
\begin{thm}
For an infinite set $U$ there exist $2^{2^{\card U}}$ equivalence
classes of isomorphic ultrafilters.\end{thm}
\begin{proof}
The number of bijections between any two given subsets of $U$ is
no more than $(\card U)^{\card U}=2^{\card U}$. The number of bijections
between all pairs of subsets of $U$ is no more than $2^{\card U}\cdot2^{\card U}=2^{\card U}$.
Therefore each isomorphism class contains at most $2^{\card U}$ ultrafilters.
But there are $2^{2^{\card U}}$ ultrafilters. So there are $2^{2^{\card U}}$
classes.\end{proof}
\begin{rem}
One of the above mentioned equivalence classes contains trivial ultrafilters.\end{rem}
\begin{cor}
There exist non-isomorphic nontrivial ultrafilters on any infinite
set.
\end{cor}
\section{Consequences}
\begin{thm}
\label{triv-atom-prod}The graph of reloid $\mathcal{F}\times^{\mathsf{RLD}}\uparrow^{A}\{a\}$
is isomorphic to the filter $\mathcal{F}$ for every set $A$ and
$a\in A$.\end{thm}
\begin{proof}
From \ref{inj-iso-dom}.\end{proof}
\begin{thm}
If $f$, $g$ are reloids, $f\sqsubseteq g$ and $g$ is monovalued
then $g|_{\dom f}=f$.\end{thm}
\begin{proof}
It's simple to show that $f=\bigsqcup\setcond{f|_{a}}{a\in\atoms^{\mathscr{F}(\Src f)}}$
(use the fact that $k\sqsubseteq f|_{a}$ for some $a\in\atoms^{\mathscr{F}(\Src f)}$
for every $k\in\atoms f$ and the fact that $\mathsf{RLD}(\Src f,\Dst f)$
is atomistic).
Suppose that $g|_{\dom f}\neq f$. Then there exists $a\in\atoms\dom f$
such that $g|_{a}\neq f|_{a}$.
Obviously $g|_{a}\sqsupseteq f|_{a}$.
If $g|_{a}\sqsupset f|_{a}$ then $g|_{a}$ is not atomic (because
$f|_{a}\ne\bot^{\mathsf{RLD}(\Src f,\Dst f)}$) what contradicts to
a theorem above. So $g|_{a}=f|_{a}$ what is a contradiction and thus
$g|_{\dom f}=f$.\end{proof}
\begin{cor}
\label{mv-is-restr}Every monovalued reloid is a restricted principal
monovalued reloid.\end{cor}
\begin{proof}
Let $f$ be a monovalued reloid. Then there exists a function $F\in\GR f$.
So we have
\[
(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}F)|_{\dom f}=f.
\]
\end{proof}
\begin{cor}
Every monovalued injective reloid is a restricted injective monovalued
principal reloid.\end{cor}
\begin{proof}
Let $f$ be a monovalued injective reloid. There exists a function
$F$ such that $f=(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}F)|_{\dom f}$.
Also there exists an injection $G\in\up f$.
Thus
\begin{multline*}
f=f\sqcap(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}G)|_{\dom f}=\\
(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}F)|_{\dom f}\sqcap(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}G)|_{\dom f}=\\
(\uparrow^{\mathsf{RLD}(\Src f,\Dst f)}(F\sqcap G))|_{\dom f}.
\end{multline*}
Obviously $F\sqcap G$ is an injection.\end{proof}
\begin{thm}
If a reloid $f$ is monovalued and $\dom f$ is an principal filter
then $f$ is principal.\end{thm}
\begin{proof}
$f$ is a restricted principal monovalued reloid. Thus $f=F|_{\dom f}$