this repo has no description
dekker.one
1% RUNS ON mzn20_fd
2% RUNS ON mzn-fzn_fd
3% RUNS ON mzn20_fd_linear
4% RUNS ON mzn20_mip
5
6% Regression test for a bug in mzn2fzn 1.2. The optimisation pass was leaving
7% dangling references to variables it had "eliminated". The symptom was the
8% following error from the FlatZinc interpreter:
9%
10% subsets_100.fzn:413:
11% symbol error: `INT____407' undeclared
12%
13% (This model is from the original bug report.)
14
15% Subsets 100 puzzle in MiniZinc.
16%
17% From rec.puzzle FAQ
18% http://brainyplanet.com/index.php/Subsets?PHPSESSID=051ae1e2b6df794a5a08fc7b5ecf8028
19% """
20% Out of the set of integers 1,...,100 you are given ten different integers.
21% From this set, A, of ten integers you can always find two disjoint non-empty
22% subsets, S & T, such that the sum of elements in S equals the sum of elements
23% in T. Note: S union T need not be all ten elements of A. Prove this.
24% """
25
26%
27% This MiniZinc model was created by Hakan Kjellerstrand, hakank@bonetmail.com
28% See also my MiniZinc page: http://www.hakank.org/minizinc
29%
30
31include "globals.mzn";
32int: n = 100;
33int: m = 10;
34var set of 1..n: s;
35var set of 1..n: t;
36var int: s_total;
37var int: t_total;
38
39%
40% sums the integer in set ss
41%
42predicate sum_set(var set of int: ss, var int: total) =
43 let {
44 int: m = card(ub(ss)),
45 array[1..m] of var 0..1: tmp
46 }
47 in
48 forall(i in 1..m) (
49 i in ss <-> tmp[i] = 1
50 )
51 /\
52 total = sum(i in 1..m) (i*tmp[i])
53;
54
55
56solve :: set_search([s,t],
57 input_order, indomain_min, complete) satisfy;
58
59constraint
60 card(s union t) <= m
61 /\
62 card(s union t) > 0
63 /\
64 disjoint(s, t)
65 /\
66 sum_set(s, s_total)
67 /\
68 sum_set(t, t_total)
69 /\
70 s_total = t_total
71 % /\
72 % t_total = n
73;
74
75output [
76 "s: ", show(s), "\n",
77 "t: ", show(t), "\n",
78 "s_total: ", show(s_total), "\n",
79 "t_total: ", show(t_total), "\n",
80];