| blob | 40be562b69462a602dc5169276cd396059bb05ce |
1 version: 0.28
2 date: Thu Oct 16 21:56:01 CEST 2008
3 changes:
4 - add omega as submodule
5 - minor compilation and bug fixes
6 ---
7 version: 0.27
8 date: Sun May 4 23:42:49 CEST 2008
9 changes:
10 - add Parker's library for DFA based counting
11 - support nested sums based exact summation
12 - support Laurent expansion based exact summation
13 - handle equalities and factors in summation
14 - support range propagation for computing bounds
15 - support computation of exact bound through iteration
16 - fix conversion from GiNaC's numeric to Values
17 ---
18 version: 0.26
19 date: Wed Jan 16 23:45:17 CET 2008
20 changes:
21 - support computation of generating functions in barvinok_enumerate_e
22 - support polymake 2.3
23 - support GiNaC 1.3.8 and newer
24 - support summation using Bernoulli formula
25 - add and use zsolve from 4ti2 for computing Hilbert bases of cones
26 - support use of piplib for solving LPs
27 - add polyhedron_integer_hull and polytope_minimize
28 - add polytope_lattice_width for computing the lattice widths of polytopes
29 ---
30 version: 0.25
31 date: Tue Sep 11 22:02:22 CEST 2007
32 changes:
33 - add PolyLib and piplib as submodules
34 - remove compatibility with old versions of PolyLib and piplib
35 - support use of piplib for basis reduction
36 - support summation using local Euler-Maclaurin formula for 2D polytopes
37 - support TOPCOM based chamber decompositions
38 - simplify primal decomposition
39 ---
40 version: 0.24
41 date: Sat Jun 30 23:37:31 CEST 2007
42 changes:
43 - support exponential (Todd) substitution for non-parametric problems
44 - support primal and stopped decomposition on parametric problems
45 - support nested sums polynomial approximation of quasi-polynomials
46 - add barvinok_summate for summating a polynomial over a polytope
47 - speed improvements in barvinok_count and polyhedron_sample
48 ---
49 version: 0.23
50 date: Sat Apr 28 20:10:35 CEST 2007
51 changes:
52 - use arpg for argument parsing in most commands
53 - |-
54 support polynomial approximations of quasi-polynomials, including
55 volume computation, Meister's method and replacing fractional parts
56 by constants
57 - add barvinok_maximize for computing upper (and lower) bounds
58 - support non-parametric primal and stopped Barvinok decomposition
59 - periodic number representation now run-time configurable
60 - drop support for ancient PolyLib versions
61 - simplified partitioning of chamber complex
62 - add evalue_eval, an exact version of copmute_evalue
63 - improved handling of large generating functions
64 - integrate barvinok_series functionality into barvinok_enumerate
65 - support use of cdd for solving LPs in basis reduction
66 - initial support for computing bounds over quasi-polynomials
67 - recursive Bernstein expansion
68 - corrections and improvements of lexmin
69 - optionally print quasi-polynomial as list of polynomials
70 - add some documentation
71 ---
72 version: 0.22
73 date: Tue Oct 31 23:15:15 CET 2006
74 changes:
75 - include bernstein library with interface to barvinok and occ
76 - some occ clean-ups
77 - barvinok_count now only returns -1 for unbounded polyhedra with integer points
78 - preliminary implementation of projection via neighborhood complexes
79 - implement generalized basis reduction
80 - add polyhedron_sample
81 - add polytope_scan
82 - add lexmin
83 - some internal reorganization
84 - some gen_fun clean-ups and optimizations
85 - plug some memory leaks
86 - fix some corner cases
87 - shared libraries are no longer built by default
88 - introduce barvinok_options to pass to functions instead of just MaxRays
89 - use same LLL reduction parameter that LattE uses
90 - add some documentation
91 ---
92 version: 0.21
93 date: Thu Jun 22 20:49:11 CEST 2006
94 changes:
95 - add some documentation
96 - add barvinok_ehrhart
97 - add h_star_vector polymake client
98 - fix some corner cases
99 - more experimental support for enumerating unions of polytopes
100 - rename cdd2polylib.pl to latte2polylib.pl
101 ---
102 version: 0.20
103 date: Wed Feb 22 12:01:23 CET 2006
104 changes:
105 - fix PIP-based reduction by removing some code and assuming a recent PIP
106 - some memory leaks plugged
107 - optionally use Omega as a preprocessor in barvinok_enumerate_e
108 - integrate verif_ehrhart_e into barvinok_enumerate_e
109 - barvinok_series handles more problem with equalities
110 - add gen_fun::Hadamard_product
111 - experimental support for enumerating unions of polytopes
112 ---
113 version: 0.19
114 date: Fri Jan 20 22:14:44 CET 2006
115 changes:
116 - include Omega Calculator interface occ
117 - print_enode creates parameter names if none are provided
118 - only use PIP to project out some variables
119 - some header file fixes
120 - C89 fixes again
121 ---
122 version: 0.18
123 date: Thu Sep 15 11:27:19 CEST 2005
124 changes:
125 - factor polyhedra into independent parts before enumeration
126 - make "random vector" specialization more robust
127 - speed up verif_ehrhart by only checking inside the context
128 - remove {Param,}Polyhedron_Reduce
129 - rename triangularize_cone to triangulate_cone
130 - generalize computation of lexmin/max using PIP
131 - compute the number of lexsmaller elements using PolyLib's LexSmaller
132 ---
133 version: 0.17
134 date: Mon Aug 1 12:23:50 CEST 2005
135 changes:
136 - fix some configuration problems
137 - include minimal polymake client
138 ---
139 version: 0.16
140 date: Thu Jun 30 12:10:29 CEST 2005
141 changes:
142 - implement "breadth-first" incremental specialization
143 - implement non-incremental version for series computation
144 - make non-incremental specialization the default again
145 - check for bad PolyLib version at runtime
146 - compiles on Windows (mingw32)
147 - add version option
148 - allow crosscompilation
149 - compute the coefficients of an Ehrhart series
150 - fix conversion of PIP output to polyhedra
151 - allow direct verification of Ehrhart series
152 ---
153 version: 0.15
154 date: Sun Dec 12 23:00:40 CET 2004
155 changes:
156 - export Polyhedron_Project
157 - allow verification of step-functions computed from Ehrhart series
158 - optionally compute step-functions incrementally
159 - disable possibly incorrect optimization
160 - minor optimizations in Ehrhart series computations
161 ---
162 version: 0.14
163 date: Wed Nov 17 19:22:23 CET 2004
164 changes:
165 - use code for computing series to count non-parametric polytopes
166 - add conversion from series to explicit function
167 - fix removal of redundant equalities
168 ---
169 version: 0.13
170 date: Sat Oct 16 12:13:30 CEST 2004
171 changes:
172 - use slightly larger random numbers
173 - change basis to remove existential variable
174 - fix bug in reduction
175 - handle empty PIP solution
176 - add utility to remove equalities
177 ---
178 version: 0.12
179 date: Sun Oct 3 20:47:11 CEST 2004
180 changes:
181 - reduce memory requirements
182 - add cdd2polylib.pl
183 - add patch for NTL 5.3.2
184 - experimental support for Ehrhart series
185 - compile fix on systems that lack <getopt.h>
186 ---
187 version: 0.11
188 date: Sat Sep 18 14:49:40 CEST 2004
189 changes:
190 - fix minor bugs
191 - find and use PIP if available
192 - don't overallocate matrices on newer PolyLibs
193 - verification program for enumeration of integer projections
194 - summation of enumerations
195 - avoid gccisms
196 - support older versions of GMP
197 ---
198 version: 0.10
199 date: Tue Jul 6 22:09:11 CEST 2004
200 changes:
201 - install header files
202 - modulo representation changed to fractionals
203 - perform some simplifications on the modulo representation
204 - barvinok_enumerate returns a malloc'ed evalue
205 - a few bugs in barvinok_enumerate_e removed
206 - optionally print an indication of the size of an enumeration
207 - emask added
208 - bug in triangularization fixed
209 - bug in polyhedron reduction fixed
210 - verif_ehrhart.c should compile on C89 systems again
211 ---
212 version: 0.09
213 changes:
214 - bug in polyhedron reduction fixed
215 ---
216 version: 0.08
217 date: Sun Feb 22 21:15:00 CET 2004
218 changes:
219 - remove remaining part of Kristof's Ehrhart manipulation code
220 - optimize modulo expression based on validity domain
221 - support partition evalues
222 - factor out 1D polytopes
223 - validity domains can now be unions of polytopes
224 - optionally convert modulo representation to lookup-table
225 - rename indicator to relation
226 - add DomainIncludes
227 ---
228 version: 0.07
229 changes:
230 - support "indicator" evalues
231 - support equalities for modulo representation
232 ---
233 version: 0.06
234 changes:
235 - remove part of Kristof's Ehrhart manipulation code
236 - partial support for modulo representation
237 - support cvs version of PolyLib
238 ---
239 version: 0.05
240 date: Thu Nov 27 13:35:59 CET 2003
241 changes:
242 - |-
243 include code from Kristof Beyls for manipulating Ehrhart
244 polynomials
245 - |-
246 implement barvinok_enumerate for constructing Ehrhart
247 polynomials using barvinok's algorithm
248 ---
249 version: 0.04
250 date: Thu Nov 6 01:36:25 CET 2003
251 changes:
252 - |-
253 allow specification of max #constraints
254 for those without automatically growing chernikova tables
255 - fix a few small bugs
256 - some minor optimizations
257 ---
258 version: 0.03
259 date: Mon Nov 3 16:35:57 CET 2003
260 changes:
261 - fix a number of bugs
262 - add tests
263 ---
264 version: 0.02
265 date: Mon Oct 20 14:15:45 CEST 2003
266 changes:
267 - change name from count to barvinok_count
268 - return -1 if polyhedron is unbounded
269 - trivially optimize counting in presence of some independent constraints