1.1 --- /dev/null Thu Jan 01 00:00:00 1970 +0000
1.2 +++ b/ChangeLog Mon Dec 06 13:09:21 2010 +0100
1.3 @@ -0,0 +1,2356 @@
1.4 +Sun Dec 05 12:00:00 2010 Andrew Makhorin <mao@gnu.org>
1.5 +
1.6 + * GLPK 4.45 (30:0:30) has been released
1.7 +
1.8 + * glplpx01.c
1.9 + A bug (it_cnt) in routine reset_parms was fixed.
1.10 + Thanks to Ali Baharev <ali.baharev@gmail.com> for report.
1.11 +
1.12 + * glpmpl03.c
1.13 + A bug (print "text\") was fixed.
1.14 + Thanks to Xypron <xypron.glpk@gmx.de> for report.
1.15 +
1.16 + * glpsql.c
1.17 + A precision bug was fixed.
1.18 + Thanks to Xypron <xypron.glpk@gmx.de>.
1.19 +
1.20 + * glpk.tex
1.21 + Some typos were corrected.
1.22 + Thanks to Robbie Morrison <robbie@actrix.co.nz>.
1.23 +
1.24 +Thu Jun 03 12:00:00 2010 Andrew Makhorin <mao@gnu.org>
1.25 +
1.26 + * GLPK 4.44 (29:0:29) has been released
1.27 +
1.28 + * glpapi14.c glpmpl.h glpmpl01.c glpmpl03.c glpmpl04.c
1.29 + Implemented suffixes for variables and constraints.
1.30 +
1.31 + * glpmpl06.c
1.32 + Made changes to allow comment records in CSV files.
1.33 +
1.34 + * glpapi17.c
1.35 + Added and documented new API routine glp_cpp to solve Critical
1.36 + Path Problem.
1.37 +
1.38 +Sat Feb 20 12:00:00 2010 Andrew Makhorin <mao@gnu.org>
1.39 +
1.40 + * GLPK 4.43 (28:0:28) has been released
1.41 +
1.42 + * glplib.h, glplib.c, glpenv.h, glpenv.c
1.43 + The module glpenv was split into two modules glpenv and glplib.
1.44 +
1.45 + * glpenv01.c, glpenv03.c, glpenv04.c, glpenv06.c
1.46 + The following new API routines were added and documented:
1.47 + glp_init_env, glp_free_env, glp_open_tee, glp_close_tee,
1.48 + glp_error (macro), glp_difftime.
1.49 +
1.50 + * glpapi16.c
1.51 + New API routine glp_top_sort (topological sorting of ayclic
1.52 + digraph) was added and documented.
1.53 +
1.54 + * glpapi17.c
1.55 + A serious bug was fixed in the routine glp_asn_prob_hall.
1.56 +
1.57 + * glpnpp05.c
1.58 + A bug was fixed in the LP/MIP preprocessor (hidden covering
1.59 + inequalities).
1.60 +
1.61 + * glpsql.c
1.62 + Some improvements were made in the table driver (NULL data).
1.63 + Thanks to Xypron <xypron.glpk@gmx.de> for contribution.
1.64 +
1.65 + * configure.ac
1.66 + Changes were made to use .dylib rather than .so under Mac OS.
1.67 + Thanks to Noli Sicad <nsicad@gmail.com> for testing
1.68 +
1.69 +Wed Jan 13 12:00:00 2010 Andrew Makhorin <mao@gnu.org>
1.70 +
1.71 + * GLPK 4.42 (27:0:27) has been released
1.72 +
1.73 + * glpapi01.c, glpapi11.c, glpapi12.c, glpdmx.c
1.74 + The following new API routines were added and documented:
1.75 + glp_check_dup (check for duplicate elements in sparse matrix);
1.76 + glp_sort_matrix (sort elements of the constraint matrix);
1.77 + glp_read_prob (read problem data in GLPK format);
1.78 + glp_write_prob (write problem data in GLPK format);
1.79 + glp_analyze_bound (analyze active bound of non-basic variable);
1.80 + glp_analyze_coef (analyze obj. coefficient at basic variable);
1.81 + glp_print_ranges (print sensitivity analysis report; replaces
1.82 + lpx_print_sens_bnds).
1.83 +
1.84 + * glpapi20.c
1.85 + New command-line options were added to glpsol:
1.86 + --glp (read problem data in GLPK format);
1.87 + --wglp (write problem data in GLPK format);
1.88 + --lp (replaces --cpxlp);
1.89 + --wlp (replaces --wcpxlp);
1.90 + --ranges (print sensitivity analysis report).
1.91 +
1.92 + * glpapi06.c
1.93 + In the routine glp_init_smcp default value of the parameter
1.94 + out_frq was changed to 500 (it was 200).
1.95 +
1.96 + * glpipp.h, glpipp01.c, glpipp02.c
1.97 + The old MIP preprocessor module was removed.
1.98 +
1.99 + * glpapi09.c
1.100 + Now the MIP solver uses the new MIP preprocessor (NPP).
1.101 +
1.102 + * glplpx03.c
1.103 + lpx_write_opb was disabled due to replacing IPP with NPP.
1.104 +
1.105 + * glpnet09.c
1.106 + Kellerman's heuristic to cover edges by cliques was added.
1.107 +
1.108 + * glplib08.c
1.109 + Recognition of special filenames "/dev/stdin", "/dev/stdout",
1.110 + and "/dev/stderr" was added.
1.111 +
1.112 + * glpk.tex
1.113 + Chapter "Graph and network routines" was carried out from the
1.114 + reference manual as a separate document.
1.115 +
1.116 +Mon Dec 21 12:00:00 2009 Andrew Makhorin <mao@gnu.org>
1.117 +
1.118 + * GLPK 4.41 (26:0:26) has been released
1.119 +
1.120 + * glpapi12.c
1.121 + The following new API routines were added:
1.122 + glp_transform_row (replaces lpx_transform_row);
1.123 + glp_transform_col (replaces lpx_transform_col);
1.124 + glp_prim_rtest (replaces lpx_prim_ratio_test);
1.125 + glp_dual_rtest (replaces lpx_dual_ratio_test).
1.126 + Note that values returned by glp_prim_rtest and glp_dual_rtest
1.127 + differ from the ones retutned by the deprecated routines.
1.128 +
1.129 + * glpnpp*.*
1.130 + The LP/MIP preprocessor was essentially re-implemented.
1.131 +
1.132 + * glpios03.c
1.133 + The feature to remove inactive cuts from the active subproblem
1.134 + was implemented.
1.135 +
1.136 + * glpios11.c
1.137 + The feature processing cuts stored in the cut pool was improved
1.138 + (now it uses estimation of objective degradation).
1.139 +
1.140 + * glpscg.*
1.141 + Obsolete implemetation of the conflict graph was removed.
1.142 +
1.143 + * glpmpl.h, glpmpl03.c, glpmpl04.c
1.144 + FILE was replaced by XFILE to allow using GLPK I/O routines.
1.145 +
1.146 + * glpsql.c, examples/sql, doc/tables.tex
1.147 + The SQL table driver was changed to allow multiple arguments
1.148 + separated by semicolon in SQL statements. Thanks to Xypron
1.149 + <xypron.glpk@gmx.de>.
1.150 +
1.151 + * glpk.h, glpapi14.c
1.152 + New API routine glp_time was added (not documented yet).
1.153 +
1.154 + * glpapi20.c
1.155 + Two new options were added to glpsol: --seed value (initialize
1.156 + pseudo-random number generator used in MathProg model with
1.157 + specified seed value), and --ini filename (use as initial basis
1.158 + previously saved with -w option).
1.159 +
1.160 + * examples/xyacfs.mod
1.161 + Thanks to Nigel Galloway <nigel_galloway@operamail.com> for
1.162 + contribution.
1.163 +
1.164 + * examples/dbf/*.*
1.165 + Thanks to Noli Sicad <nsicad@gmail.com> for contribution.
1.166 +
1.167 + * w32/*.*, w64/*.*
1.168 + Scripts to build GLPK with Microsoft Visual Studio 2010 were
1.169 + added. Thanks to Xypron <xypron.glpk@gmx.de> for contribution
1.170 + and testing.
1.171 +
1.172 +Tue Nov 03 12:00:00 2009 Andrew Makhorin <mao@gnu.org>
1.173 +
1.174 + * GLPK 4.40 (25:0:25) has been released
1.175 +
1.176 + * glpdmx.c
1.177 + Two new API routines were added:
1.178 + glp_read_ccdata (read graph in DIMACS clique/coloring format);
1.179 + glp_write_ccdata (write graph in DIMACS clique/coloring format).
1.180 + Also an example file examples/sample.col was added.
1.181 +
1.182 + * glpapi19.c, glpnet08.c
1.183 + New API routine glp_wclique_exact was added. It is intended to
1.184 + find a maximum weight clique with the exact algorithm developed
1.185 + by Prof. P. Ostergard.
1.186 +
1.187 + * glpnpp02.c
1.188 + A bug was fixed in the LP preprocessor (routine npp_empty_col).
1.189 + Thanks to Stefan Vigerske <stefan@math.hu-berlin.de> for the
1.190 + bug report.
1.191 +
1.192 + * glpios10.c
1.193 + A bug was fixed and some improvements were made in the FPUMP
1.194 + heuristic module. Thanks to Xypron <xypron.glpk@gmx.de>.
1.195 +
1.196 + * glpapi12.c
1.197 + A bug was fixed in the API routine glp_warm_up (dual
1.198 + feasibility test was incorrect in maximization case). Thanks to
1.199 + Uday Venkatadri <Uday.Venkatadri@dal.ca> for the bug report.
1.200 +
1.201 + * glpapi16.c
1.202 + Two new API routines were added:
1.203 + glp_del_vertices (remove vertices from graph);
1.204 + glp_del_arc (remove arc from graph).
1.205 +
1.206 + * glpios09.c
1.207 + The hybrid pseudocost branching heuristic was included in the
1.208 + MIP solver. It is available on API level (iocp.br_tech should
1.209 + be set to GLP_BR_PCH) and in the stand-alone solver glpsol
1.210 + (via the command-line option --pcost). This heuristic may be
1.211 + useful on solving hard MIP instances.
1.212 +
1.213 + * glpios03.c
1.214 + The branching heuristic by Driebeck and Tomlin (used in the
1.215 + MIP solver by default) was changed to switch to branching on
1.216 + most fractional variable if an lower bound of degradation of
1.217 + the objective is close to zero for all branching candidates.
1.218 +
1.219 +Sun Jul 26 12:00:00 2009 Andrew Makhorin <mao@mai2.rcnet.ru>
1.220 +
1.221 + * GLPK 4.39 (24:0:24) has been released
1.222 +
1.223 + * glpsdf.c
1.224 + New API routines to read plain data files were added.
1.225 +
1.226 + * glpcpx.h, glpini.h, glpscl.h
1.227 + These headers were removed.
1.228 +
1.229 + * glpcpx.c
1.230 + API routines glp_read_lp and glp_write_lp to read/write files
1.231 + in CPLEX LP format were re-implemented. Now glp_write_lp
1.232 + correctly writes double-bounded (ranged) rows by introducing
1.233 + slack variables rather than by duplicating the rows. The data
1.234 + structure glp_cpxcp and routine glp_init_cpxcp were added.
1.235 +
1.236 + * amd/*
1.237 + The 'xfree(NULL)' bug was fixed in the AMD routines. Thanks to
1.238 + Niels Klitgord <niels@bu.edu> for the bug report.
1.239 +
1.240 + * glpapi16.c
1.241 + New API routines glp_set_vertex_name, glp_create_v_index,
1.242 + glp_find_vertex, and glp_delete_v_index were added.
1.243 +
1.244 + * glpdmx.c
1.245 + New API routines glp_read_asnprob, glp_write_asnprob,
1.246 + glp_read_ccformat, and glp_write_ccformat were added (the two
1.247 + latter routines are not documented yet).
1.248 +
1.249 + * glpapi18.c
1.250 + New API routines glp_check_asnprob, glp_asnprob_lp,
1.251 + glp_asnprob_okalg, and glp_asnprob_hall were added.
1.252 +
1.253 + * glpini01.c, glpini02.c
1.254 + The message "Crashing..." was changed to "Constructing initial
1.255 + basis..." due to suggestion by Thomas Kahle <tom111@gmx.de>.
1.256 +
1.257 + * glpapi14.c
1.258 + New API routines glp_printf, glp_vprintf, glp_malloc,
1.259 + glp_calloc, glp_free, and glp_assert were added.
1.260 +
1.261 + * glplpp.h, glplpp01.c, glplpp02.c
1.262 + Old LP presolver routines were removed. Now glp_simplex uses
1.263 + new preprocessing routines (see glpnpp).
1.264 +
1.265 + * glpapi12.c
1.266 + New API routine glp_warm_up was added; it replaces the routine
1.267 + lpx_warm_up.
1.268 +
1.269 +Sat May 02 12:00:00 2009 Andrew Makhorin <mao@mai2.rcnet.ru>
1.270 +
1.271 + * GLPK 4.38 (23:0:23) has been released
1.272 +
1.273 + * glpmps.c
1.274 + API routines to read/write MPS files were re-implemented.
1.275 +
1.276 + * glpspx02.c
1.277 + Some improvements were made in the dual simplex routine.
1.278 +
1.279 + * glpk.h
1.280 + New structure glp_iptcp was added.
1.281 +
1.282 + * glpnpp.h, glpnpp01.c, glpnpp02.c
1.283 + New LP/MIP preprocessor. Currently it includes only some basic
1.284 + routines and used only in the interior-point solver.
1.285 +
1.286 + * glpapi08.c
1.287 + API routine glp_interior was replaced by an improved version
1.288 + (new LP/MIP preprocessor, new ordering algorithms).
1.289 +
1.290 + New API routine glp_init_iptcp was added.
1.291 +
1.292 + API routine glp_ipt_status may return two new statuses due to
1.293 + changes in glp_interior.
1.294 +
1.295 + * glpsol.c
1.296 + New command-line options were added (ordering algorithm used in
1.297 + the interior-point solver).
1.298 +
1.299 + * amd/*.*, colamd/*.*
1.300 + Two external software modules AMD and COLAMD/SYMAMD used in the
1.301 + interior-point solver were included in the distribution.
1.302 +
1.303 + For details see amd/README and colamd/README.
1.304 +
1.305 + * glpnet03.c, glpnet04.c, glpnet05.c
1.306 + A minor bug was fixed (_G => G_). Thanks to Nelson H. F. Beebe
1.307 + <beebe@math.utah.edu> for bug report.
1.308 +
1.309 +Sun Mar 29 12:00:00 2009 Andrew Makhorin <mao@mai2.rcnet.ru>
1.310 +
1.311 + * GLPK 4.37 (22:0:22) has been released
1.312 +
1.313 + * glpk.h
1.314 + iocp.fp_heur was added to enable/disable fpump heuristic.
1.315 +
1.316 + * glpios10.c
1.317 + ios_feas_pump was added (feasibility pump heuristic).
1.318 +
1.319 + * glpsol.c
1.320 + --fpump command-line option was added.
1.321 +
1.322 + * glpsds.c
1.323 + Plain data set routines were added to facilitate reading plain
1.324 + data in application programs. Currently these routines are not
1.325 + in API, though declared in glpk.h.
1.326 +
1.327 + * glpapi08.c
1.328 + A bug was fixed in the internal routine restore. (Due to this
1.329 + bug dual solution components were computed incorrectly if the
1.330 + problem was scaled.)
1.331 +
1.332 + * glpapi10.c, glpapi11.c
1.333 + The following new API routines were added:
1.334 + glp_print_sol (replaces lpx_print_sol);
1.335 + glp_print_ipt (replaces lpx_print_ips);
1.336 + glp_print_mip (replaces lpx_print_mip);
1.337 + _glp_check_kkt (replaces lpx_check_kkt, lpx_check_int).
1.338 + Now the routine lpx_print_prob (deprecated) is equivalent to
1.339 + the routine glp_write_lp.
1.340 +
1.341 + * glpapi18.c, glpapi19.c
1.342 + The following new API routines were added:
1.343 + glp_read_graph (read graph from plain text file);
1.344 + glp_write_graph (write graph to plain text file);
1.345 + glp_weak_comp (find all weakly connected components);
1.346 + glp_strong_comp (find all strongly connected components).
1.347 +
1.348 + * configure.ac, Makefile.am
1.349 + Changes were made: (a) to allow using autoreconf/autoheader;
1.350 + (b) to allow building glpk in a directory other than its source
1.351 + directory. Thanks to Marco Atzeri <marco_atzeri@yahoo.it> for
1.352 + bug report.
1.353 +
1.354 + * examples/shiftcover.mod
1.355 + An example model in MathProg was added.
1.356 + Thanks to Larry D'Agostino <Larry.D'Agostino@gmacrescap.com>
1.357 + for contribution.
1.358 +
1.359 +Fri Feb 06 12:00:00 2009 Andrew Makhorin <mao@mai2.rcnet.ru>
1.360 +
1.361 + * GLPK 4.36 (21:0:21) has been released
1.362 +
1.363 + * glpnet06.c, glpnet07.c, glpapi19.c
1.364 + The following new API routines were added:
1.365 + glp_mincost_okalg find minimum-cost flow with out-of-kilter
1.366 + algorithm
1.367 + glp_maxflow_ffalg find maximal flow with Ford-Fulkerson
1.368 + algorithm
1.369 +
1.370 + * glpsol.c
1.371 + Two new command-line options were added:
1.372 + --mincost read min-cost flow data in DIMACS format
1.373 + --maxflow read maximum flow data in DIMACS format
1.374 +
1.375 + * doc/glpk.*
1.376 + New edition of the reference manual was included.
1.377 +
1.378 + * glpk.h
1.379 + Duplicate symbols were removed to allow using swig.
1.380 + Thanks to Kelly Westbrooks <kellywestbrooks@yahoo.com> and
1.381 + Nigel Galloway <nigel_galloway@operamail.com> for suggestion.
1.382 +
1.383 + * glpcpx.c
1.384 + A minor defect was fixed in the routine glp_write_lp.
1.385 + Thanks to Sebastien Briais <sbriais@free.fr> for bug report.
1.386 +
1.387 + * glpsql.c
1.388 + A minor bug was fixed. Thanks to Xypron <xypron.glpk@gmx.de>
1.389 + for patch.
1.390 +
1.391 + * examples/hashi.mod, examples/shikaku.mod
1.392 + Two example models in MathProg were added. Thanks to Sebastian
1.393 + Nowozin <nowozin@gmail.com> for contribution.
1.394 +
1.395 + * examples/qfit.mod, examples/yacfs.mod
1.396 + Two example models in MathProg were added. Thanks to Nigel
1.397 + Galloway <nigel_galloway@operamail.com> for contribution.
1.398 +
1.399 +Fri Jan 09 12:00:00 2009 Andrew Makhorin <mao@mai2.rcnet.ru>
1.400 +
1.401 + * GLPK 4.35 (20:0:20) has been released
1.402 +
1.403 + * glpk.h, glpapi.c, glpnet.c
1.404 + The following new API routines were added:
1.405 + glp_create_graph create graph
1.406 + glp_set_graph_name assign (change) graph name
1.407 + glp_add_vertices add new vertices to graph
1.408 + glp_add_arc add new arc to graph
1.409 + glp_erase_graph erase graph content
1.410 + glp_delete_graph delete graph
1.411 + glp_read_mincost read minimum cost flow problem data in
1.412 + DIMACS format
1.413 + glp_write_mincost write minimum cost flow problem data in
1.414 + DIMACS format
1.415 + glp_mincost_lp convert minimum cost flow problem to LP
1.416 + glp_netgen Klingman's network problem generator
1.417 + glp_gridgen grid-like network problem generator
1.418 + glp_read_maxflow read maximum flow problem data in DIMACS
1.419 + format
1.420 + glp_write_maxflow write maximum flow problem data in DIMACS
1.421 + format
1.422 + glp_maxflow_lp convert maximum flow problem to LP
1.423 + glp_rmfgen Goldfarb's maximum flow problem generator
1.424 +
1.425 + * doc/glpk.*
1.426 + New edition of the reference manual was included.
1.427 +
1.428 + * examples/sample.min, examples/sample.max
1.429 + Two example data files in DIMACS format were added.
1.430 +
1.431 + * glplib04.c
1.432 + The statement "if (c = '\n') fflush(stdout)" was added to the
1.433 + internal routine xputc to provide "real-time" terminal output.
1.434 + Thanks to Luiz Bettoni <bettoni@cpgei.ct.utfpr.edu.br> for
1.435 + suggestion.
1.436 +
1.437 + * glpmpl05.c
1.438 + A minor bug was fixed in the internal routine mpl_fn_time2str.
1.439 + Thanks to Stefan Vigerske <stefan@vigerske.de> for bug report.
1.440 +
1.441 + * w32/makefile, w64/makefile
1.442 + The flag -O2 (/O2) was added to some makefiles.
1.443 +
1.444 +Thu Dec 04 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.445 +
1.446 + * GLPK 4.34 (19:0:19) has been released
1.447 +
1.448 + * src/glpios03.c
1.449 + A bug was fixed in the internal routine branch_on. Thanks to
1.450 + Nigel Galloway <nigel_galloway@operamail.com> for bug report.
1.451 +
1.452 + * src/glpmpl05.c
1.453 + Three new MathProg functions were included:
1.454 + gmtime obtaining current calendar time
1.455 + str2time converting character string to calendar time
1.456 + time2str converting calendar time to character string
1.457 + Thanks to Xypron <xypron.glpk@gmx.de>.
1.458 +
1.459 + * doc/glpk.*, doc/gmpl.*
1.460 + A new edition of the GLPK reference manual and GNU MathProg
1.461 + language description were included.
1.462 +
1.463 +Thu Oct 30 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.464 +
1.465 + * GLPK 4.33 (18:0:18) has been released
1.466 +
1.467 + * glpapi*.*
1.468 + The following new API routines were added:
1.469 + glp_copy_prob copy problem object content
1.470 + glp_exact solve LP in exact arithmetic
1.471 + (makes lpx_exact deprecated)
1.472 + glp_get_unbnd_ray determine variable causing unboundedness
1.473 + (makes lpx_get_ray_info deprecated)
1.474 + glp_interior solve LP with interior-point method
1.475 + (makes lpx_interior deprecated)
1.476 +
1.477 + * glpapi*.*
1.478 + The following new API routines for processing models written in
1.479 + the GNU Mathprog language were added to the package:
1.480 + glp_mpl_alloc_wksp allocate the translator workspace
1.481 + glp_mpl_read_model read and translate model section
1.482 + glp_mpl_read_data read and translate data section
1.483 + glp_mpl_generate generate the model
1.484 + glp_mpl_build_prob build LP/MIP instance from the model
1.485 + glp_mpl_postsolve postsolve the model
1.486 + glp_mpl_free_wksp deallocate the translator workspace
1.487 + (These routines make lpx_read_model deprecated.)
1.488 +
1.489 + * src/glpapi17.c, examples/glpsol.c
1.490 + The stand-alone solver glpsol was re-implemented with new API
1.491 + routines.
1.492 +
1.493 + * src/glpsql.c
1.494 + Some bugs were fixed in the SQL table driver. Thanks to Xypron
1.495 + <xypron.glpk@gmx.de>.
1.496 +
1.497 + * examples/cplex/*.*
1.498 + A crude implementation of CPLEX-like interface to GLPK API was
1.499 + added to the package. See examples/cplex/README.
1.500 +
1.501 +Fri Oct 03 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.502 +
1.503 + * GLPK 4.32 (17:0:17) has been released
1.504 +
1.505 + * glpmpl01.c
1.506 + A bug was fixed. Due to this bug iterated expressions having
1.507 + an indexing expression whose dummy indices are bound to some
1.508 + values, i.e. like sum{(i+1,j,k-1) in E} x[i,j,k] are evaluated
1.509 + incorrectly. Namely, current value of such expressions is not
1.510 + invalidated when corresponding dummy indices (like i and k in
1.511 + the example above) are changed, that erroneously results in the
1.512 + same value evaluated for the first time.
1.513 +
1.514 + * glpios03.c
1.515 + Euclidean reduction of the local objective bound was added in
1.516 + the routine glpios03.c.
1.517 +
1.518 + * glpapi11.c
1.519 + The following new branch-and-cut API routines were added:
1.520 + glp_ios_row_attr determine additional row attributes;
1.521 + glp_ios_pool_size determine current size of the cut pool;
1.522 + glp_ios_add_row add constraint to the cut pool;
1.523 + glp_ios_del_row delete constraint from the cut pool;
1.524 + glp_ios_clear_pool delete all constraints from the cut pool.
1.525 +
1.526 + * glpapi08.c
1.527 + The following new features were included in the branch-and-cut
1.528 + solver (the API routine glp_intopt):
1.529 + MIP presolver;
1.530 + mixed cover cut generator;
1.531 + clique cut generator.
1.532 + Due to the MIP presolver glp_intopt may additionally return
1.533 + GLP_ENOPFS and GLP_ENODFS, if primal or dual infeasibility of
1.534 + LP relaxation is detected by the presolver. Also the return
1.535 + code GLP_EMIPGAP was introduced to correctly indicate that the
1.536 + mip gap tolerance is reached.
1.537 +
1.538 + * glplpx01.c
1.539 + Now the obsolete API routines lpx_integer and lpx_intopt are
1.540 + completely superseded by the API routine glp_intopt that makes
1.541 + them deprecated.
1.542 +
1.543 + * glpmpl05.c
1.544 + Now the table driver name "iODBC" can be specified as "ODBC".
1.545 +
1.546 + * glpmpl03.c
1.547 + A bug fixed in the routine free_dca.
1.548 + Thanks to Xypron <xypron.glpk@gmx.de>.
1.549 +
1.550 + * glpsql.c
1.551 + A bug was fixed in the SQL table driver.
1.552 + Thanks to Xypron <xypron.glpk@gmx.de>.
1.553 +
1.554 + * examples/glpsol.c
1.555 + Changes were made to allow multiple MathProg data files.
1.556 +
1.557 + * doc/glpk.*
1.558 + A new edition of the GLPK reference manual was included.
1.559 +
1.560 + * doc/tables.*
1.561 + A new edition of the supplement "Using data tables in the GNU
1.562 + MathProg language" was included.
1.563 +
1.564 +Tue Sep 02 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.565 +
1.566 + * GLPK 4.31 (16:0:16) has been released
1.567 +
1.568 + * glpspx.h, glpspx01.c, glpspx02.c, glpapi06.c
1.569 + The dual simplex solver (spx_dual_opt) was replaced by a new
1.570 + implementation of the two-phase dual simplex method (spx_dual).
1.571 + Old simplex method routines (spx_prim_opt, spx_prim_feas, and
1.572 + spx_dual_opt) were removed from the package.
1.573 +
1.574 + * glpk.h, glpscl.h, glpscl.c, glpapi04.c
1.575 + New API routine glp_scale_prob was added. It replaces routine
1.576 + lpx_scale_prob which is deprecated.
1.577 +
1.578 + * glpk.h, glpini.h, glpini01.c, glpini02.c, glpapi05.c
1.579 + New API routines glp_std_basis, glp_adv_basis, glp_cpx_basis
1.580 + were added. They replace routines lpx_std_basis, lpx_adv_basis,
1.581 + lpx_cpx_basis which are deprecated.
1.582 +
1.583 + * glpdmp.c
1.584 + 8-byte data alignment was added to the module (sufficient for
1.585 + both ILP32 and LP64 environments).
1.586 +
1.587 + * glplib07.c
1.588 + 16-byte data alignment was added to the module to provide
1.589 + compatibility with LP64 environment (8-byte is not sufficient
1.590 + due to jmp_buf; thanks to Xypron for investigation).
1.591 +
1.592 + * glplpx16.c
1.593 + New version of the routine lpx_write_pb was added. Thanks to
1.594 + Oscar Gustafsson <oscarg@isy.liu.se> for the contribution.
1.595 +
1.596 + * w32/VC9, w64/VC9
1.597 + Makefiles and batch files were added to build GLPK under 32-
1.598 + and 64-bit Windows with Microsoft Visual Studio Express 2008.
1.599 + Thanks to Heinrich Schuchardt <heinrich.schuchardt@gmx.de> for
1.600 + the contribution and testing.
1.601 +
1.602 + * w32/DM, w32/OWC
1.603 + Makefiles and batch files were added to build GLPK with Digital
1.604 + Mars C/C++ 8.50 and Open Watcom C/C++ 1.6 (32-bit Windows).
1.605 +
1.606 +Wed Aug 13 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.607 +
1.608 + * GLPK 4.30 (15:0:15) has been released
1.609 +
1.610 + * glpspx.h, glpspx03.c, glpapi06.c
1.611 + The primal simplex solver (spx_prim_opt, spx_prim_feas) was
1.612 + replaced by a new implementation (spx_primal), which currently
1.613 + provides the same features as the old version.
1.614 +
1.615 + * glpmpl01.c, glpmpl03.c
1.616 + Some changes were made in the MathProg translator to allow <,
1.617 + <=, >=, and > on comparing symbolic values. Thanks to Heinrich
1.618 + Schuchardt <heinrich.schuchardt@gmx.de> for patches.
1.619 +
1.620 + * glplpx10.c
1.621 + Internal routine set_d_eps in the exact LP solver was changed
1.622 + to prevent approximation errors in case of integral data.
1.623 + Thanks to Markus Pilz <pilz@cs.uni-bonn.de> for bug report.
1.624 +
1.625 +XXX XXX XX 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.626 +
1.627 + * GLPK 4.29 (14:0:14) has been released
1.628 +
1.629 + * configure.ac
1.630 + The configure script was changed to disable optional features
1.631 + by default. For details see file INSTALL.
1.632 +
1.633 + * glpipp02.c
1.634 + A bug was fixed in the internal routine reduce_bounds. Thanks
1.635 + to Anne-Laurence Putz <anne-laurence.putz@eurodecision.com> for
1.636 + the bug report.
1.637 +
1.638 + * glpapi01.c
1.639 + New API routine glp_erase_prob was added.
1.640 +
1.641 + * glpapi13.c
1.642 + New API routines glp_read_mps and glp_write_mps were added.
1.643 + They replace API routines lpx_read_mps, lpx_read_freemps,
1.644 + lpx_write_mps, and lpx_write_freemps, which are deprecated.
1.645 +
1.646 + * glpapi14.c
1.647 + New API routines glp_read_lp and glp_write_lp were added. They
1.648 + replace API routines lpx_read_cpxlp and lpx_write_cpxlp, which
1.649 + are deprecated.
1.650 +
1.651 + * glpsql.c
1.652 + Minor bug was fixed. Thanks to Xypron <xypron.glpk@gmx.de> for
1.653 + the bug report.
1.654 +
1.655 +Tue Mar 25 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.656 +
1.657 + * GLPK 4.28 (13:0:13) has been released
1.658 +
1.659 + * glplib.h, glplib.c
1.660 + Three wrapper routines xdlopen, xdlsym, and xdlclose, which
1.661 + provide the shared library support, were added. A particular
1.662 + version of these routines depends on the option --enable-dl
1.663 + passed to the configure script (see file INSTALL for details).
1.664 + Thanks to Rafael Laboissiere <rafael@debian.org> for useful
1.665 + advices concerning the shared library support.
1.666 +
1.667 + * glpsql.c
1.668 + A static linking to iODBC and MySQL libraries used in the
1.669 + MathProg table drivers was replaced by a dynamic linking to
1.670 + corresponding shared libraries.
1.671 + Many thanks to Heinrich Schuchardt <heinrich.schuchardt@gmx.de>
1.672 + for the contribution and to Vijay Patil <vijay.patil@gmail.com>
1.673 + for testing this feature under Windows XP.
1.674 +
1.675 + * glpgmp.h, glpgmp.c
1.676 + A bug (which appeared only on 64-bit platforms) was fixed.
1.677 + Thanks to Axel Simon <Axel.Simon@ens.fr> for the bug report.
1.678 +
1.679 + * glpapi.c
1.680 + A bug was fixed in the api routine glp_add_cols. (If the basis
1.681 + is valid, adding column keeps it valid, however, col->bind was
1.682 + set to -1 rather to 0.)
1.683 + Thanks to Cedric[FR] <fox2113@wanadoo.fr> for the bug report.
1.684 +
1.685 + * glplib.c
1.686 + 64-bit unsigned int type glp_ulong and corresponding routines
1.687 + were replaced by 64-bit signed int type xlong_t.
1.688 +
1.689 + * glpk.h, glpapi.c
1.690 + The type glp_ulong was replaced by glp_long. This affects only
1.691 + the api routine glp_mem_usage.
1.692 +
1.693 + * glplib.c
1.694 + Compressed data file support was added. This feature requires
1.695 + the zlib data compression libraries and allows compressing and
1.696 + decompressing .gz files "on the fly".
1.697 +
1.698 + * glpcli.h, glpcli.c
1.699 + Command-line interface routines were added. (This feature is
1.700 + incomplete so far.)
1.701 +
1.702 +Sun Mar 02 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.703 +
1.704 + * GLPK 4.27 (12:0:12) has been released
1.705 +
1.706 + * glpsql.h, glpsql.c
1.707 + Two MathProg table drivers for iODBC and MySQL contributed by
1.708 + Heinrich Schuchardt <heinrich.schuchardt@gmx.de> were added to
1.709 + the package.
1.710 +
1.711 + * glpmpl05.c
1.712 + Mathprog table driver for xBASE was added to the package.
1.713 +
1.714 + * glpmpl03.c
1.715 + A minor was fixed in the MathProg translator (if some field
1.716 + specified in the table statement is missing in corresponding
1.717 + input table, the bug causes abnormal termination). Thanks to
1.718 + Heinrich Schuchardt <heinrich.schuchardt@gmx.de> for the bug
1.719 + report.
1.720 +
1.721 + * glpmpl.h, glpmpl.c
1.722 + STRING data type was replaced by plain character strings.
1.723 +
1.724 +Sun Feb 17 12:00:00 2008 Andrew Makhorin <mao@mai2.rcnet.ru>
1.725 +
1.726 + * GLPK 4.26 (11:0:11) has been released
1.727 +
1.728 + * glpmpl.h, glpmpl01.c, glpmpl03.c, glpmpl05.c
1.729 + The table statement was implemented. Description of this new
1.730 + feature is given in file doc/tables.txt.
1.731 +
1.732 + * glpios03.c
1.733 + A bug causing zero divide error on computing euclidean norm of
1.734 + the cut coefficient vector was fixed.
1.735 +
1.736 +Wed Dec 19 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.737 +
1.738 + * GLPK 4.25 (10:0:10) has been released
1.739 +
1.740 + * glpapi10.c
1.741 + Routines lpx_eval_tab_row and lpx_eval_tab_col were replaced by
1.742 + glp_eval_tab_row and glp_eval_tab_col.
1.743 +
1.744 + * glpios03.c, glpios05.c
1.745 + Gomory's mixed integer cuts were implemented.
1.746 +
1.747 + * glpscs.h, glpscs.c
1.748 + Segmented character string routines are no longer used and were
1.749 + removed from the package.
1.750 +
1.751 +Wed Nov 21 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.752 +
1.753 + * GLPK 4.24 (9:0:9) has been released
1.754 +
1.755 + * src/glplpx16.c
1.756 + A bug was fixed in the routine lpx_write_cpxlp. If a variable
1.757 + x has upper bound and no lower bound, it should appear in the
1.758 + bounds section as "-inf <= x <= u", not as "x <= u". Thanks to
1.759 + Enric Rodriguez <erodri@lsi.upc.edu> for the bug report.
1.760 +
1.761 + * src/glpios03.c, src/glpios04.c, src/glpios05.c
1.762 + MIR (mixed integer rounding) cuts were implemented.
1.763 +
1.764 +Sun Oct 28 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.765 +
1.766 + * GLPK 4.23 (8:0:8) has been released
1.767 +
1.768 + * src/glplib05.c, configure.ac
1.769 + Check for vsnprintf was added.
1.770 +
1.771 + * include/glppds.h, src/glppds.c
1.772 + A module to scan plain data was added.
1.773 +
1.774 + * src/glpapi09.c
1.775 + The following new API routines were added:
1.776 + glp_read_sol read basic solution from text file;
1.777 + glp_write_sol write basic solution to text file;
1.778 + glp_read_ipt read interior-point solution from text file;
1.779 + glp_write_ipt write interior-point solution to text file;
1.780 + glp_read_mip read MIP solution from text file;
1.781 + glp_write_mip write MIP solution to text file.
1.782 +
1.783 + * src/glpapi12.c
1.784 + Advanced API routine glp_free_env was added.
1.785 +
1.786 + * examples/glpsol.c
1.787 + The following three command-line options were added:
1.788 + --mipgap tol set relative MIP gap tolerance
1.789 + -r filename read solution from filename
1.790 + -w filename write solution to filename
1.791 +
1.792 +Wed Sep 19 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.793 +
1.794 + * GLPK 4.22 (7:0:7) has been released
1.795 +
1.796 + * src/glpios02.c
1.797 + A bug was fixed in the MIP preprocessor (ios_preprocess_node).
1.798 + Thanks to Roberto Bagnara <bagnara@cs.unipr.it> (Department of
1.799 + Mathematics, University of Parma, Italy) for the bug report.
1.800 +
1.801 + * src/glpios02.c
1.802 + A bug was fixed in the MIP preprocessor (col_implied_bounds),
1.803 + due to which constraint coefficients with small magnitude could
1.804 + lead to wrong implied bounds of structural variables.
1.805 +
1.806 + * src/glpipp02.c
1.807 + A similar bug was fixed in the routine reduce_bounds.
1.808 +
1.809 + * src/glpapi01.c
1.810 + A bug was fixed in the routines glp_set_mat_row and
1.811 + glp_set_mat_col. (The bug appeared due to incorrect removing
1.812 + zero elements from the row/column lists.)
1.813 +
1.814 + * src/glplpx14.c
1.815 + A bug was fixed in the API routines lpx_read_mps and
1.816 + lpx_read_freemps, due to which bounds of type LI specified in
1.817 + BOUNDS section were incorrectly processed.
1.818 +
1.819 + * src/glplib05.c
1.820 + A call to standard function vsprintf was replaced by a call to
1.821 + vsnprintf for security reasons. Many thanks to Peter T. Breuer
1.822 + <ptb@inv.it.uc3m.es> and Rafael Laboissiere <rafael@debian.org>.
1.823 +
1.824 +Tue Aug 28 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.825 +
1.826 + * GLPK 4.21 (6:0:6) has been released
1.827 +
1.828 + * glpscg.h, glpscg.c
1.829 + Routines to maintain sparse cliqued graph were added.
1.830 +
1.831 + * glpios02.c
1.832 + MIP preprocessing routines were added.
1.833 +
1.834 + * glpk.h, glpios.h, glpios03.c
1.835 + New reasons for calling the callback routine were introduced
1.836 + in the MIP solver.
1.837 +
1.838 + * glpapi08.c
1.839 + Default backtracking strategy was changed to best local bound.
1.840 +
1.841 + * glpapi11.c
1.842 + New API routine glp_term_out to enable/disable terminal output
1.843 + was added.
1.844 +
1.845 + * glprng.h, glprng02.c
1.846 + Two routines to generate uniformly distributed pseudo-random
1.847 + floating-point numbers were added.
1.848 +
1.849 +Thu Jul 26 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.850 +
1.851 + * GLPK 4.20 (5:0:5) has been released
1.852 +
1.853 + * glpk.h, glpapi08.c
1.854 + The routine lpx_integer was replaced by an equivalent routine
1.855 + glp_intopt. Also new API routine glp_init_iocp was added.
1.856 +
1.857 + * glpiet.h, glpiet.c
1.858 + Routines implementing the implicit enumeration tree are
1.859 + no longer used and therefore were removed from the package.
1.860 +
1.861 + * glpios.h, glpios01.c, glpios02, glpios03
1.862 + Routines implementing the integer optimization suite being
1.863 + replaced by a new version were removed from the package.
1.864 +
1.865 + * glpmip.h, glpmip01.c, glpmip02.c
1.866 +
1.867 + Routines implementing the B&B method being replaced by a new
1.868 + version were removed from the package.
1.869 +
1.870 + * glpios.h, glpios01.c, glpios02.c
1.871 +
1.872 + Routines implementing a new version of the integer optimization
1.873 + suite (IOS) based on the B&B method were added to the package.
1.874 +
1.875 + * glpk.h, glpapi10.c
1.876 + Branch-and-bound interface routines were added to the package.
1.877 +
1.878 + * examples/tspsol.c
1.879 + The TSP solver based on old version of the integer optimization
1.880 + suite is no more supported and was removed from the package.
1.881 +
1.882 + * glpipp02.c
1.883 + An error in the routine reduce_bounds was fixed; thanks to
1.884 + Graham Rockwell <bionomicron@gmail.com> for the bug report.
1.885 +
1.886 + * glpk.latex
1.887 + A new edition of the reference manual was included.
1.888 +
1.889 +Thu Jul 05 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.890 +
1.891 + * GLPK 4.19 (4:0:4) has been released
1.892 +
1.893 + The principal change is upgrading to GPLv3.
1.894 +
1.895 + * glpapi01.c
1.896 + A serious bug in the routine glp_del_cols was fixed; thanks to
1.897 + Cedric[FR] <fox2113@wanadoo.fr> for the bug report. The bug
1.898 + appeared because on deleting non-basic columns the basis header
1.899 + remained valid, however, contained invalid (old) column ordinal
1.900 + numbers.
1.901 +
1.902 + * glpapi10.c
1.903 + A new advanced API routine glp_mem_limit was added.
1.904 +
1.905 + * glplpx01.c
1.906 + The case GLP_EBOUND was added to the routine lpx_simplex.
1.907 + Thanks to Cameron Kellough <Cameron.Kellough@sri.com> for the
1.908 + bug report.
1.909 +
1.910 + * glplpx19.c
1.911 + An API routine lpx_write_pb to write the problem instance in
1.912 + OPB (pseudo boolean) format format was added. Thanks to Oscar
1.913 + Gustafsson <oscarg@isy.liu.se> for the contribution.
1.914 +
1.915 + * glpsol.c
1.916 + Two new options --wpb and --wnpb were added to glpsol to write
1.917 + the problem instance in OPB format.
1.918 +
1.919 +Mon Jun 25 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.920 +
1.921 + * GLPK 4.18 (3:0:3) has been released
1.922 +
1.923 + * glplib.h
1.924 + Type names ulong_t and uldiv_t were changed to glp_ulong and
1.925 + glp_uldiv to avoid conflicts with standard type names on some
1.926 + platforms. Thanks to Boris Wirtz <Boris.Wirtz@uni-oldenburg.de>
1.927 + for the bug report.
1.928 +
1.929 + * glpbfd.*, glpfhv.*, glplpf.*
1.930 + LP basis factorization routines were made tidy.
1.931 +
1.932 + * glpk.h, glpapi04.c
1.933 + The following API routines were added:
1.934 + glp_set_rii, glp_set_sjj, glp_get_rii, glp_get_sjj.
1.935 +
1.936 + * glpk.h, glpapi06.c
1.937 + The routine lpx_simplex was replaced by an equivalent routine
1.938 + glp_simplex. Also new API routine glp_init_smcp was added.
1.939 +
1.940 + * glpk.h, glpapi09.c
1.941 + The following advanced API routines were added:
1.942 + glp_bf_exists, glp_factorize, glp_bf_updated, glp_get_bfcp,
1.943 + glp_set_bfcp, glp_get_bhead, glp_get_row_bind, glp_get_col_bind,
1.944 + glp_ftran, glp_btran.
1.945 +
1.946 + * glpk.latex
1.947 + A new edition of the reference manual was included.
1.948 +
1.949 + * examples/dea.mod, examples/food.mod, examples/food2.mod
1.950 + Three examples in the MathProg language were added.
1.951 + Thanks to Sebastian Nowozin <nowozin@gmail.com>.
1.952 +
1.953 +Sat May 26 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.954 +
1.955 + * GLPK 4.17 (2:0:2) has been released
1.956 +
1.957 + * glpdmp.h, glpdmp.c
1.958 + Memory pool routines were replaced by a new version.
1.959 +
1.960 + * glpscs.h, glpscs.c
1.961 + Segmented string routines were replaced by a new version.
1.962 +
1.963 + * glplpx08.c, glplpx09.c
1.964 + Now the MIP problem may have no integer columns.
1.965 +
1.966 + * glpapi01.c
1.967 + The routines glp_set_mat_row, glp_set_mat_col, and glp_load_mat
1.968 + were modified to allow zero elements (which are not stored in
1.969 + the constraint matrix).
1.970 +
1.971 + * glpscf.h, glpscf.c
1.972 + Schur complement factorization routines were implemented.
1.973 +
1.974 + * glplpf.h, glplpf.c
1.975 + LP basis factorization routines based on LU-factorization and
1.976 + Schur complement were implemented.
1.977 +
1.978 + * glplpx02.c, glplpx03.c
1.979 + New control parameter LPX_K_BFTYPE was introduced to choose the
1.980 + basis factorization type used by the simplex method routines.
1.981 +
1.982 + * glpsol.c
1.983 + Three new command-line options were added to choose the basis
1.984 + factorization type used by the simplex method routines: --luf,
1.985 + --cbg, and --cgr.
1.986 +
1.987 + * glpk.latex
1.988 + A new edition of the reference manual was included.
1.989 +
1.990 +Sat May 05 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.991 +
1.992 + * GLPK 4.16 (1:0:1) has been released
1.993 +
1.994 + * glpk.h, glpapi.c, glplpx01.c, glplpx02.c
1.995 + Names of a number basic api routines were changed and now have
1.996 + the prefix 'glp_'. To keep backward compatibility these routines
1.997 + are also available via their old names prefixed with 'lpx_'.
1.998 +
1.999 + * glplpx19.c
1.1000 + Three new api routines were added: glp_version, glp_term_hook,
1.1001 + and glp_mem_usage.
1.1002 +
1.1003 + * glpk.latex, gmpl.texi
1.1004 + A new edition of the reference manuals was included.
1.1005 +
1.1006 + * lpglpk40.c
1.1007 + This example program is no longer supported and therefore was
1.1008 + removed from the package.
1.1009 +
1.1010 +Sun Feb 18 12:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1011 +
1.1012 + * GLPK 4.15 (0:0:0) has been released
1.1013 +
1.1014 + * configure.ac, Makefile.am
1.1015 + Autotools specification files were changed to use GNU Libtool
1.1016 + that allows building the static as well as shared GLPK library.
1.1017 + Thanks to Rafael Laboissiere <rafael@debian.org>.
1.1018 +
1.1019 +Mon Feb 05 08:00:00 2007 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1020 +
1.1021 + * GLPK 4.14 has been released
1.1022 + Now GLPK conforms to ILP32, LLP64, and LP64 programming models
1.1023 + (the latter seems to be the ultimate choice regarding 64-bit
1.1024 + architectures). Note that GLPK itself is a 32-bit application,
1.1025 + and the conformity only means that the package works correctly
1.1026 + on all these arenae. Nevertheless, on 64-bit platforms it is
1.1027 + possible to use more than 4GB of memory, if necessary.
1.1028 +
1.1029 + * Makefile
1.1030 + Starting from this release only the header glpk.h is needed to
1.1031 + be installed.
1.1032 +
1.1033 + * glplib01.c
1.1034 + Two routines bigmul and bigdiv which performs multiplication
1.1035 + and division of unsigned integers of arbitrary precision were
1.1036 + added.
1.1037 +
1.1038 + * glplib02.c
1.1039 + A set of 64-bit arithmetic routines were added.
1.1040 +
1.1041 + * glplib04.c
1.1042 + Some low-level library routines were improved and renamed.
1.1043 +
1.1044 + * glpcfg.h
1.1045 + The macro GLP_TM_SPEC were introduced to specify a version of
1.1046 + the time routine depending on the host environment.
1.1047 +
1.1048 +Mon Nov 13 12:00:00 2006 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1049 +
1.1050 + * GLPK 4.13 has been released
1.1051 +
1.1052 + * configure.in
1.1053 + '-lm' bug was fixed.
1.1054 +
1.1055 + * glpbfx.h, glpbfx.c
1.1056 + Basis factorization interface routines based on exact (bignum)
1.1057 + arithmetic were implemented.
1.1058 +
1.1059 + * glpssx.h, glpssx1.c, glpssx2.c
1.1060 + Simplex method routines based on exact (bignum) arithmetic were
1.1061 + implemented.
1.1062 +
1.1063 + * glplpx6e.c
1.1064 + The routine lpx_exact, which is an easy-to-use driver to the
1.1065 + exact simplex method, was added.
1.1066 +
1.1067 + * glpsol.c
1.1068 + Two command-line options were added: '--exact' and '--xcheck'.
1.1069 +
1.1070 +Wed Nov 08 12:00:00 2006 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1071 +
1.1072 + * GLPK 4.12 has been released
1.1073 +
1.1074 + * glpcfg.h
1.1075 + The package configuration file was added.
1.1076 +
1.1077 + * glplib2.c
1.1078 + Alternative version of the routines umalloc, ucalloc, and ufree
1.1079 + was provided. It does not limit the amount of allocated memory
1.1080 + to INT_MAX bytes and therefore can be used on platforms where
1.1081 + sizeof(void *) > sizeof(int). To enable this version one should
1.1082 + define the preprocessor variable GLP_HUGE_MEM.
1.1083 +
1.1084 + * glprng.c
1.1085 + The routine rng_create_rand was changed to initialize the
1.1086 + generator using seed = 1, not 0, to conform ISO C requirements.
1.1087 +
1.1088 + * glpgmp.h, glpgmp.c
1.1089 + A set of bignum arithmetic routines implementing operations on
1.1090 + integers and rationals was added. These routines are compatible
1.1091 + with the GNU MP library.
1.1092 +
1.1093 + NOTE: To attain a much better performance it is recommended to
1.1094 + use, if possible, the original GNU MP library rather than the
1.1095 + GLPK version, by defining the preprocessor variable GLP_USE_GMP.
1.1096 +
1.1097 + * glplux.h, glplux.c
1.1098 + A tentative implementation of sparse LU-factorization based on
1.1099 + exact (bignum) arithmetic was added.
1.1100 +
1.1101 + * glpssx.h, glpssx.c
1.1102 + A tentative implementation of some simplex method routines based
1.1103 + on exact (bignum) arithmetic was added.
1.1104 +
1.1105 + * glplpx6f.c
1.1106 + A preliminary implementation of the routine lpx_exact_check was
1.1107 + added. This routine checks the current basis for primal and dual
1.1108 + feasibility using exact (bignum) arithmetic.
1.1109 +
1.1110 + * examples/glpsol.c
1.1111 + The command-line option '--xcheck' was introduced to check the
1.1112 + current basis for feasibility using exact (bignum) arithmetic.
1.1113 +
1.1114 +Tue Jul 25 12:00:00 2006 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1115 +
1.1116 + * GLPK 4.11 has been released.
1.1117 +
1.1118 + * include/glpbfi.h, src/glpbfi.c
1.1119 + Basis factorization interface routines were added.
1.1120 +
1.1121 + * include/glpluf.h, src/glpluf1.c
1.1122 + Hypersparse solution routines were added.
1.1123 +
1.1124 + * include/glpinv.h, src/glpinv1.c
1.1125 + Hypersparse solution routines (fake version) were added.
1.1126 +
1.1127 + * include/glpmpl.h, src/glpmpl.c
1.1128 + Built-in functions card, length, and substr were implemented.
1.1129 + Output redirection in the printf statement was implemented.
1.1130 +
1.1131 + * examples/graph.mod, examples/crypto.mod
1.1132 + Two example models illustrating new features of the modeling
1.1133 + language were included.
1.1134 +
1.1135 +Thu May 11 12:00:00 2006 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1136 +
1.1137 + * GLPK 4.10 has been released.
1.1138 +
1.1139 + * src/glplpx8a.c
1.1140 + A fragment was added to the routines lpx_read_mps and
1.1141 + lpx_read_freemps to accept LI bound type (it is similar to LO,
1.1142 + however, additionally marks the column as integer).
1.1143 +
1.1144 + * include/glpbfi.h, src/glpbfi.c
1.1145 + The module glpbfi which implements the basis factorization
1.1146 + interface (BFI) was added.
1.1147 +
1.1148 + * src/glplpx7a.c
1.1149 + The routine lpx_cover_cut to generate mixed cover cuts was
1.1150 + added.
1.1151 +
1.1152 + * src/glplpx7b.c
1.1153 + The routine lpx_clique_cut to generate clique cuts and related
1.1154 + routines to maintain the conflict graph were added.
1.1155 +
1.1156 + * include/glplpx.h, src/glplpx5.c
1.1157 + The routine lpx_cpx_basis implementing Bixby's algorithm to
1.1158 + construct an initial LP basis was added.
1.1159 +
1.1160 + * examples/glpsol.c
1.1161 + Command-line option '--bib' was added which allows building
1.1162 + an initial LP basis using Bixby's algorithm.
1.1163 + Default command-line option '--mps' was changed to '--freemps'.
1.1164 +
1.1165 + * examples/cf12a.mod, examples/cf12b.mod
1.1166 + Two examples in MathProg (curve fitting problem) were added.
1.1167 + Thanks to Dr. Harley Mackenzie <hjm@hardsoftware.com>.
1.1168 +
1.1169 +Tue Jan 17 12:00:00 2006 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1170 +
1.1171 + * GLPK 4.9 has been released.
1.1172 +
1.1173 + * glpipp.h, glpipp1.c, glpipp2.c
1.1174 + A MIP presolver were implemented (currently incomplete). It is
1.1175 + used internally in the routine lpx_intopt (see below).
1.1176 +
1.1177 + * glplpx6d.c, glplpx7a.c
1.1178 + An advanced branch-and-bound solver (the routine lpx_intopt)
1.1179 + were implemented.
1.1180 +
1.1181 + * glplpx6c.c
1.1182 + The routine lpx_check_int to check MIP feasibility conditions
1.1183 + was added.
1.1184 +
1.1185 + * glplpx8a.c
1.1186 + The routine lpx_print_mip was changed to print MIP feasibility
1.1187 + conditions.
1.1188 +
1.1189 + * glpmpl.h, glpmpl1.c, glpmpl3.c
1.1190 + The built-in functions sin, cos, atan, and atan2 were added to
1.1191 + the MathProg language.
1.1192 +
1.1193 + * doc/lang.*
1.1194 + Some typos were fixed.
1.1195 + Thanks to Minh Ha Duong <haduong@centre-cired.fr> (CIRED, CNRS).
1.1196 +
1.1197 +Wed Jan 12 12:00:00 2005 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1198 +
1.1199 + * GLPK 4.8 has been released.
1.1200 +
1.1201 + * glpspx.h, glpspx1.c, glpspx2.c, glplpx6a.c
1.1202 + Simplex method routines were changed due to a new format of the
1.1203 + constraint matrix.
1.1204 +
1.1205 + * glpmat.h, glpmat.c
1.1206 + Sparse matrix routines were re-implemented using storage-by-rows
1.1207 + format.
1.1208 +
1.1209 + * glpipm.h, glpipm.c, glplpx6b.c
1.1210 + Interior-point method routines were changed due to a new format
1.1211 + of sparse matrices.
1.1212 +
1.1213 + * glpchol.h, glpchol.c
1.1214 + Old version of Cholesky factorization routines being replaced by
1.1215 + a new one (see glpmat.c) was removed from the package.
1.1216 +
1.1217 + * glplpx8c.c
1.1218 + Minor bug was fixed in api routine lpx_read_cpxlp.
1.1219 +
1.1220 +Mon Aug 23 12:00:00 2004 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1221 +
1.1222 + * GLPK 4.7 has been released.
1.1223 +
1.1224 + * glplpx.h, glplpx1.c
1.1225 + New core API routines were added (but not documented yet):
1.1226 + lpx_order_matrix, lpx_create_index, lpx_find_row, lpx_find_col,
1.1227 + lpx_delete_index.
1.1228 +
1.1229 + * glplpx8a.c
1.1230 + API routine lpx_read_mps was re-implemented, and two new API
1.1231 + routines lpx_read_freemps and lpx_write_freemps were added to
1.1232 + support free MPS format.
1.1233 +
1.1234 + * glplpx8c.c
1.1235 + Two API routines lpx_read_cpxlp and lpx_write_cpxlp (formerly
1.1236 + named lpx_read_lpt and lpx_write_lpt) were re-implemented.
1.1237 +
1.1238 + * glpmps.h, glpmps.c
1.1239 + This module formerly used in lpx_read_mps was removed from the
1.1240 + package.
1.1241 +
1.1242 + * glplpt.h, glplpt.c
1.1243 + This module formerly used in lpx_read_lpt was removed from the
1.1244 + package.
1.1245 +
1.1246 + * glpmip.h, glpmip1.h, glpmip2.h
1.1247 + New MIP routines mip_best_node and mip_relative_gap were added
1.1248 + due to suggestion of Brady Hunsaker <hunsaker@engr.pitt.edu>.
1.1249 +
1.1250 + * glpsol.c
1.1251 + The following new command-options were added:
1.1252 + --freemps to read problem data in free MPS format
1.1253 + --wfreemps to write problem data in free MPS format
1.1254 + --cpxlp to read problem data in CPLEX LP format
1.1255 + --wcpxlp to write problem data in CPLEX LP format
1.1256 + --bas to read LP basis from a text file in MPS format
1.1257 + --wbas to write LP basis to a text file in MPS format
1.1258 + --mostf to use "most fractional" branching heuristic
1.1259 + --bestb to use "best bound" backtracking heuristic
1.1260 +
1.1261 + * contrib/deli/*.*
1.1262 + GLPK Delphi interface module was temporarily removed from the
1.1263 + distribution due to licensing problems.
1.1264 +
1.1265 + * contrib/glpkmex/*.*
1.1266 + GLPK Matlab interface module was temporarily removed from the
1.1267 + distribution due to licensing problems.
1.1268 +
1.1269 + * contrib/jni/*.*
1.1270 + GLPK Java interface module was temporarily removed from the
1.1271 + distribution due to licensing problems.
1.1272 +
1.1273 +Wed Aug 04 12:00:00 2004 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1274 +
1.1275 + * GLPK 4.6 has been released.
1.1276 +
1.1277 + * glpmpl.h, glpmpl1.c, glpmpl2.c, glpmpl3.c, glpmpl4.c
1.1278 + Three new statements were implemented in the GNU MathProg
1.1279 + language: solve, printf, and for. Also some bugs were fixed.
1.1280 +
1.1281 + * glplpx.h, glplpx8e.c
1.1282 + Two API routines were added: lpx_read_prob and lpx_write_prob,
1.1283 + which allow reading and writing problem data in GNU LP format.
1.1284 +
1.1285 + * glpsol.c
1.1286 + Three new command-line options were added: --glp (to read
1.1287 + problem data in GNU LP format), --wglp (to write problem data
1.1288 + in GNU LP format), and --name (to change problem name).
1.1289 +
1.1290 + * glprng.h, glprng.c
1.1291 + A portable pseudo-random number generator was implemented as a
1.1292 + separate module.
1.1293 +
1.1294 + * glplib4.c
1.1295 + The old implementation of a pseudo-random number generator was
1.1296 + removed from the package.
1.1297 +
1.1298 + * doc/lang.*, doc/refman.*
1.1299 + New edition of the GLPK documentation was included.
1.1300 +
1.1301 + * contrib/glpkmex/*.*
1.1302 + A new version of GLPKMEX was included in the distribution. For
1.1303 + more details see contrib/glpkmex/ChangeLog.
1.1304 +
1.1305 +Mon Jul 19 12:00:00 2004 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1306 +
1.1307 + * GLPK 4.5 has been released.
1.1308 +
1.1309 + * glpmip.h, glpmip1.c, glpmip2.c, glplpx6c.c
1.1310 + New implementation of the branch-and-bound method was added.
1.1311 + It replaces the old implementation, which was removed from the
1.1312 + package.
1.1313 +
1.1314 + * glpies.h, glpies1.c, glpies2.c, glpies3.c
1.1315 + Modules used in the old implementation of the branch-and-bound
1.1316 + method were removed from the package.
1.1317 +
1.1318 + * glplib2.c
1.1319 + Now if the preprocessor variable GLPHUGEMEM is defined, other
1.1320 + version of the routines umalloc, ucalloc, and ufree is used on
1.1321 + compiling the package. This allows avoiding memory allocation
1.1322 + problems on platforms where sizeof(void *) > sizeof(int), for
1.1323 + example, where addresses are 64-bit while integers are 32-bit.
1.1324 + The modification was made due to a bug report provided by Karel
1.1325 + Zimmermann <kzimm@diamant.jouy.inra.fr> and Christophe Caron
1.1326 + <caron@diamant.jouy.inra.fr>.
1.1327 +
1.1328 +Sat Jan 17 12:00:00 2004 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1329 +
1.1330 + * GLPK 4.4 has been released.
1.1331 +
1.1332 + * glplpx.h, glplpx*.c
1.1333 + All API routines were re-implemented using new data structures.
1.1334 + Some new API routines were added and some existing API routines
1.1335 + became obsolete as shown below:
1.1336 +
1.1337 + Obsolete API routine Equivalent new API routine
1.1338 + lpx_check_name (no more supported)
1.1339 + lpx_set_obj_c0 lpx_set_obj_coef
1.1340 + lpx_set_row_coef (no more supported)
1.1341 + lpx_set_col_coef lpx_set_obj_coef
1.1342 + lpx_load_mat (no more supported)
1.1343 + lpx_load_mat3 lpx_load_matrix
1.1344 + lpx_unmark_all (no more supported)
1.1345 + lpx_mark_row (no more supported)
1.1346 + lpx_mark_col (no more supported)
1.1347 + lpx_clear_mat (no more supported)
1.1348 + lpx_del_items lpx_del_rows, lpx_del_cols
1.1349 + lpx_get_row_bnds lpx_get_row_type, lpx_get_row_lb,
1.1350 + lpx_get_row_ub
1.1351 + lpx_get_col_bnds lpx_get_col_type, lpx_get_col_lb,
1.1352 + lpx_get_col_ub
1.1353 + lpx_get_obj_c0 lpx_get_obj_coef
1.1354 + lpx_get_row_coef (no more supported)
1.1355 + lpx_get_col_coef lpx_get_obj_coef
1.1356 + lpx_get_row_mark (no more supported)
1.1357 + lpx_get_col_mark (no more supported)
1.1358 + lpx_get_row_info lpx_get_row_stat, lpx_get_row_prim,
1.1359 + lpx_get_row_dual
1.1360 + lpx_get_col_info lpx_get_col_stat, lpx_get_col_prim,
1.1361 + lpx_get_col_dual
1.1362 + lpx_get_ips_stat lpx_ipt_status
1.1363 + lpx_get_ips_row lpx_ipt_row_prim, lpx_ipt_row_dual
1.1364 + lpx_get_ips_col lpx_ipt_col_prim, lpx_ipt_col_dual
1.1365 + lpx_get_ips_obj lpx_ipt_obj_val
1.1366 + lpx_get_mip_stat lpx_mip_status
1.1367 + lpx_get_mip_row lpx_mip_row_val
1.1368 + lpx_get_mip_col lpx_mip_col_val
1.1369 + lpx_get_mip_obj lpx_mip_obj_val
1.1370 +
1.1371 + Obsolete API routines were kept for backward compatibility,
1.1372 + however, they will be removed in the future.
1.1373 +
1.1374 + * doc/refman.*
1.1375 + New edition of the GLPK reference manual containing description
1.1376 + of all new API routines was included.
1.1377 +
1.1378 + * contrib/glpkmex/*.*
1.1379 + GLPKMEX, a Matlab MEX interface to GLPK package, contributed by
1.1380 + Nicolo Giorgetti <giorgetti@dii.unisi.it> was included.
1.1381 +
1.1382 + * doc/GLPK_FAQ.txt
1.1383 + GLPK FAQ contributed by Harley Mackenzie <hjm@bigpond.com> was
1.1384 + included.
1.1385 +
1.1386 +Fri Dec 12 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1387 +
1.1388 + * GLPK 4.3 has been released.
1.1389 +
1.1390 + * configure.in
1.1391 + The bug, due to which the standard math library is not linked on
1.1392 + some platforms, was fixed.
1.1393 +
1.1394 + * glpmpl3.c
1.1395 + The bug (0 ** y) was fixed in the routine fp_power.
1.1396 +
1.1397 + * glpmpl.h, glpmpl1.c, glpmpl3.c
1.1398 + Some new built-in functions (round, trunc, Irand224, Uniform01,
1.1399 + Uniform, Normal01, Normal) were added to the MathProg language.
1.1400 +
1.1401 + * glpmpl1.c
1.1402 + The MathProg syntax was changed to allow writing 'subj to'.
1.1403 +
1.1404 + * glplpx.h, glplpx1.c, glplpx2.c
1.1405 + The new api routine lpx_get_ray_info was added.
1.1406 +
1.1407 + * glplpx8a.c
1.1408 + The api routine lpx_print_sol was changed to print the number of
1.1409 + non-basic variable, which causes primal unboundness.
1.1410 +
1.1411 + * glpmps.c
1.1412 + The code was changed to avoid errors on compiling the package on
1.1413 + Mac OS X. Thanks to Andre Girard <andre@inrs-emt.uquebec.ca> for
1.1414 + the bug report.
1.1415 +
1.1416 + * doc/lang.*, doc/refman.*
1.1417 + Several typos were fixed and some new material was added in the
1.1418 + glpk documentation.
1.1419 +
1.1420 +Fri Nov 14 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1421 +
1.1422 + * GLPK 4.2 has been released.
1.1423 +
1.1424 + * glpiet.h, glpiet.c, glpios.h, glpios1.c, glpios2.c, glpios3.c
1.1425 + A preliminary implementation of the Integer Optimization Suite
1.1426 + (IOS) was included in the package. Eventually IOS will replace
1.1427 + the Implicit Enumeration Suite (IES).
1.1428 +
1.1429 + * glplpx.h, glplpx6d.c
1.1430 + A dummy version of the integer optimization routine lpx_intopt
1.1431 + was included in the package. Later this routine will replace the
1.1432 + routine lpx_integer.
1.1433 +
1.1434 + * examples/glpsol.c
1.1435 + A new command-line option --int-opt was added to the solver to
1.1436 + call lpx_intopt rather than lpx_integer.
1.1437 +
1.1438 + * glpbcs.h, glpbcs1.c, glpbcs2.c
1.1439 + Being replaced by IOS routines (see above) the Branch-and-Cut
1.1440 + Framework (BCS) routines were removed from the package.
1.1441 +
1.1442 + * examples/tspsol.c
1.1443 + Stand-alone Symmetric TSP solver was completely re-programmed
1.1444 + using IOS routines.
1.1445 +
1.1446 + * glplib.h, glplib2.c, glplib4.c
1.1447 + The random-number generator was implemented. It is based on the
1.1448 + module GB_FLIB from the Stanford GraphBase originally developed
1.1449 + by Donald Knuth.
1.1450 +
1.1451 + * glphbsm.c, glplpx8a.c, glpmps.c
1.1452 + All calls to fopen/fclose were replaced by corresponding calls
1.1453 + to ufopen/ufclose due to bug reports provided by Morten Welinder
1.1454 + <terra@gnu.org> and <jpark@sfwmd.gov>.
1.1455 +
1.1456 + * glpmps.c
1.1457 + The code was made re-entrant.
1.1458 +
1.1459 + * glplpx8b.c
1.1460 + API routine lpx_print_sens_bnds for bounds sensitivity analysis
1.1461 + contributed by Brady Hunsaker <hunsaker@engr.pitt.edu> was added
1.1462 + to the package. This feature is also available in glpsol via the
1.1463 + command-line option --bounds.
1.1464 +
1.1465 + * contrib/jni/*.*
1.1466 + New version of GLPK JNI (Java Native Interface) contributed by
1.1467 + Chris Rosebrugh <cpr@pobox.com> was added to the package.
1.1468 +
1.1469 + * contrib/deli/*.*
1.1470 + GLPK DELI (Delphi Interface) contributed by Ivo van Baren
1.1471 + <i.van.baren@freeler.nl> was added to the package.
1.1472 +
1.1473 + * glplpx3.c
1.1474 + Default method to scale the problem was changed to equilibration
1.1475 + scaling (lp->scale = 1 in lpx_reset_parms).
1.1476 +
1.1477 + * glplpx6a.c
1.1478 + Two minor (non-critical) typos were fixed due to report provided
1.1479 + by Andrew Hamilton-Wright <andrewhw@ieee.org>.
1.1480 +
1.1481 + * glplpp2.c
1.1482 + An untested case (line 941) had been tested due to bug report
1.1483 + provided by Jiri Spitz <jiri.spitz@telecom.cz>.
1.1484 +
1.1485 + * w32bc5.mak, w32vc6.mak, w32vc6d.mak, d32dmc.mak
1.1486 + Several makefiles were added to allow building GLPK library for
1.1487 + some non-GNU 32-bit platforms.
1.1488 +
1.1489 +Sat Aug 23 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1490 +
1.1491 + * GLPK 4.1 has been released.
1.1492 +
1.1493 + * glpmpl1.c, glpmpl3.c
1.1494 + Some bugs were fixed in the MathProg translator due to the bug
1.1495 + reports provided by Giles Thompson <gwpt1@cus.cam.ac.uk>:
1.1496 + conditional set expressions were incorrectly parsed;
1.1497 + dimen attribute was not set by default when a set was used
1.1498 + recursively in its own declaration;
1.1499 + logical expressions ... in if ... then ... else ... did not
1.1500 + work;
1.1501 + displaying set expressions did not free memory allocated for
1.1502 + temporary results.
1.1503 +
1.1504 + * glpmpl3.c (reduce_terms)
1.1505 + Implementation of summation of linear forms over domain was
1.1506 + improved to reduce complexity of that operation from O(n*n) to
1.1507 + O(n*log n). The improvement was made due to a report provided
1.1508 + by Sebastien de Menten <sdementen@hotmail.com>.
1.1509 +
1.1510 + * glplpx6a.c (line 1056), glpmip1.c (line 641)
1.1511 + Two minor bugs were fixed due to the bug report provided by
1.1512 + Kendall Demaree <kendal.demaree@verizon.net>.
1.1513 +
1.1514 + * glplpx.h, glplpx6a.c
1.1515 + The method of one artificial variable implemented in the routine
1.1516 + lpx_prim_art and used on the phase I in the glpk simplex solver
1.1517 + has a serious defect: for some lp instances it erroneously
1.1518 + reports that the problem has no primal feasible solution. This
1.1519 + error appears when the column of the artificial variable, which
1.1520 + enters the basis to make it primal feasible, has large
1.1521 + constraint coefficients, that leads to small reduced costs of
1.1522 + non-basic variables and premature termination of the search,
1.1523 + i.e. to wrong conclusion that the problem has no primal feasible
1.1524 + solution. To avoid this defect the routine lpx_prim_feas was
1.1525 + included. It implements the method of implicit artifical
1.1526 + variables (based on minimization of the sum of infeasibilities),
1.1527 + which is a bit slower but much more robust. The routine
1.1528 + lpx_prim_feas having the same functionality now is used instead
1.1529 + the routine lpx_prim_art.
1.1530 +
1.1531 + * glpinv.h, glpinv.c
1.1532 + The test used in the routine inv_update to detect low accuracy
1.1533 + after updating LU-factorization of the basis matrix was replaced
1.1534 + by a new, more robust test.
1.1535 +
1.1536 + * glplpx6c.c
1.1537 + Selecting an active node to be solved next in the routine
1.1538 + btrack_bestp was changed. Now, if any integer feasible solution
1.1539 + has not been found yet, the routine chooses an active node which
1.1540 + has the minimal sum of integer infeasibilities.
1.1541 +
1.1542 + * glpmip.h, glpmip1.c
1.1543 + The additional flag int_obj was included in the structure
1.1544 + MIPTREE used by the branch-and-bound. This flag is set in the
1.1545 + routine mip_create_tree and used in the routine is_better. It
1.1546 + means that the objective is integral, i.e. depends only on
1.1547 + integer variables with integer objective coefficients. The test
1.1548 + used in the routine check_integrality was also replaced by a
1.1549 + new, more reasonable one.
1.1550 +
1.1551 + * glplpx1.c
1.1552 + A minor bug was fixed in the routine lpx_check_name.
1.1553 +
1.1554 + * glpmpl.h, glpmpl4.c, glplpx8d.c
1.1555 + The flag skip_data was added to the parameter list of the
1.1556 + routine mpl_read_model. If this flag is set, the data section
1.1557 + in the model file is ignored. Corresponding change was made in
1.1558 + the routine lpx_read_model. Now, if both model and data files
1.1559 + are specified, the data section in the model file is ignored.
1.1560 +
1.1561 + * glplpx8c.c
1.1562 + A minor bug (wrong format used for writing free columns) in the
1.1563 + routine lpx_write_lpt was fixed due to the bug report provided
1.1564 + by Bernhard Schmidt <schmidt@math.uni-augsburg.de>
1.1565 +
1.1566 + * sample/glpsol.c
1.1567 + The command-line parameter --tmlim, which allows limiting the
1.1568 + solution time, was added.
1.1569 +
1.1570 + * doc/lang.*, doc/refman.*
1.1571 + New edition of the GLPK documentation was included.
1.1572 +
1.1573 + * java-binding/*.*
1.1574 + New version of the GLPK JNI (Java Native Interface) package was
1.1575 + included in the distribution.
1.1576 +
1.1577 + * sample/lpglpk40.c
1.1578 + A non-trivial example was added. It allows using GLPK as a base
1.1579 + LP solver for Concorde, a program for solving Traveling Salesman
1.1580 + Problem (TSP). For details see comments in lpglpk40.c.
1.1581 +
1.1582 + * sample/*.mod
1.1583 + Some examples of LP and MIP models written in GNU MathProg were
1.1584 + added.
1.1585 +
1.1586 +Tue May 06 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1587 +
1.1588 + * GLPK 4.0 has been released.
1.1589 +
1.1590 + * glpmpl.h, glpmpl1.c, glpmpl2.c, glpmpl3.c, glpmpl4.c
1.1591 + The model translator for the GNU MathProg modeling language was
1.1592 + implemented and included in the package.
1.1593 +
1.1594 + * glplpx.h, glplpx8d.c
1.1595 + The api routine lpx_read_model, which is an interface to the
1.1596 + MathProg translator, was included in the package.
1.1597 +
1.1598 + * glplpx.h, glplpx8a.c
1.1599 + The api routine lpx_print_prob for writing LP/MIP problem data
1.1600 + in plain text format was included in the package.
1.1601 +
1.1602 + * sample/glpsol.c
1.1603 + New version of the GLPK stand-alone LP/MIP solver that supports
1.1604 + the GNU MathProg modeling language was implemented.
1.1605 +
1.1606 + * doc/lang.latex, doc/lang.dvi, doc/lang.ps
1.1607 + The document "GLPK: Modeling Language GNU MathProg" was included
1.1608 + in the package.
1.1609 +
1.1610 + * doc/refman.latex, doc/refman.dvi, doc/refman.ps
1.1611 + New edition of the GLPK Reference Manual was included in the
1.1612 + package.
1.1613 +
1.1614 + * glplpx8c.c
1.1615 + A bug in the api routine lpx_write_lpt was fixed. Due to that
1.1616 + bug an addressing error occured in the routine if the objective
1.1617 + function has the non-zero constant term.
1.1618 +
1.1619 + * glplan.h, glplan1.c, glplan2.c, glplan3.c, glplan4.c,
1.1620 + * glplan5.c, glplan6.c, glplan7.c, glplan8.c, glplpx8b.c
1.1621 + All modules of the translator for the GLPK/L modeling language
1.1622 + were removed from the package, because GLPK/L being completely
1.1623 + superseded by GNU MathProg is no more supported.
1.1624 +
1.1625 +Tue Mar 25 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1626 +
1.1627 + * GLPK 3.3 has been released.
1.1628 +
1.1629 + * glplpp.h, glplpp1.c, glplpp2.c
1.1630 + An implementation of the built-in LP presolver was added to the
1.1631 + package.
1.1632 +
1.1633 + * glplpx.h
1.1634 + The flag presol was added to the structure LPX. This flag tells
1.1635 + the lpx_simplex whether the built-in LP presolver should be used
1.1636 + or not. By default this flag is off. Also three macros (namely
1.1637 + LPX_E_NOPFS, LPX_E_NODFS, and LPX_K_PRESOL) that concern using
1.1638 + the LP presolver were introduced.
1.1639 +
1.1640 + * glplpx3.c, glplpx6a.c
1.1641 + These modules was changed to use the built-in LP presolver.
1.1642 +
1.1643 + * sample/glpsol.c
1.1644 + Command line options --presol and --nopresol that concern using
1.1645 + the LP presolver were added to the stand-alone LP/MIP solver.
1.1646 +
1.1647 + * glplan1.c
1.1648 + This module was changed to allow declaring sets like A[1:10] in
1.1649 + the models written in the GLPK/L modeling language.
1.1650 +
1.1651 + * doc/refman.latex, doc/lang.latex
1.1652 + New editions of the documents "GLPK User's Guide" and "GLPK/L
1.1653 + Modeling Language" were included in the distribution.
1.1654 +
1.1655 + * java-binding/*.*
1.1656 + The package GLPK JNI (Java Native Interface) implementing Java
1.1657 + binding for GLPK was included in the distribution. This package
1.1658 + was developed and programmed by Yuri Victorovich <yuri@gjt.org>.
1.1659 +
1.1660 +Tue Feb 18 12:00:00 2003 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1661 +
1.1662 + * GLPK 3.2.4 has been released.
1.1663 +
1.1664 + * glplpx6b.c
1.1665 + The code was changed to allow auxiliary variables have non-zero
1.1666 + objective coefficients.
1.1667 +
1.1668 + Also a minor bug was fixed (the constant term was not considered
1.1669 + on displaying the objective function value).
1.1670 +
1.1671 + * sample/glpsol.c
1.1672 + The code was changed to fix a bug (the command-line option 'bfs'
1.1673 + was not recognized). The bug was fixed due to report provided by
1.1674 + Olivier <odwl@skynet.be>.
1.1675 +
1.1676 + * glplpt.c
1.1677 + The code was changed to fix a bug (binary variables were treated
1.1678 + erroneously as integer ones).
1.1679 +
1.1680 + * glplpx6b.c
1.1681 + The code was changed to fix a bug (variables that have no lower
1.1682 + bounds were incorrectly processed on converting to the standard
1.1683 + formulation). The bug was fixed due to report kindly provided by
1.1684 + Kjell Eikland <kjell.eikland@broadpark.no>.
1.1685 +
1.1686 +Mon Nov 11 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1687 +
1.1688 + * GLPK 3.2.3 has been released.
1.1689 +
1.1690 + * glpmip.h, glpmip1.c
1.1691 + A preliminary implementation of the branch-and-bound driver
1.1692 + based on the implicit enumeration suite (glpies) was added to
1.1693 + the package. This module is not documented yet.
1.1694 +
1.1695 + * glplpx6c.c
1.1696 + A new implementation of the api routine lpx_integer which now
1.1697 + is based on the b&b driver (see glpmip above) was included in
1.1698 + the package. This new implementation has exactly the same
1.1699 + functionality as the old version and therefore all changes are
1.1700 + transparent to the api user.
1.1701 +
1.1702 + * glpbbm.h, glpbbm.c
1.1703 + * glprsm.h, glprsm1.c, glprsm2.c
1.1704 + * glplp.h, glplp.c
1.1705 + These modules were removed from the package, because they were
1.1706 + used only in the old version of the routine lpx_integer, which
1.1707 + was replaced by the new version (see glplpx6c above).
1.1708 +
1.1709 + * glplpx.h, glplpx6a.c
1.1710 + The api routine lpx_check_kkt was included in the package and
1.1711 + its description was added in the reference manual. This routine
1.1712 + allows checking Karush-Kuhn-Tucker optimality conditions for an
1.1713 + LP solution.
1.1714 +
1.1715 + * glplpx.h, glplpx8a.c
1.1716 + Now the api routine lpx_print_sol also prints information about
1.1717 + "solution quality" obtained via the api routine lpx_check_kkt.
1.1718 +
1.1719 + * glplpx.h, glplpx8a.c
1.1720 + New api routines lpx_read_bas and lpx_write_bas were included
1.1721 + in the package and documented. The routine lpx_write_bas allows
1.1722 + writing a current basis from an LP object to a text file in the
1.1723 + MPS format. The routine lpx_read_bas allows reading a basis
1.1724 + prepared in the MPS format from a text file into an LP object.
1.1725 +
1.1726 + * glplpt.c
1.1727 + The parsing routine which reads LP problem data prepared in the
1.1728 + CPLEX LP format was modified to allow specifying lower bounds
1.1729 + of variables also in the form 'variable >= lower bound' (in the
1.1730 + bounds section). This modification was made due to a notice
1.1731 + provided by Ivan Luzzi <iluzzi@libero.it>.
1.1732 +
1.1733 + * glplpx.h, glplpx8c.c
1.1734 + The api routine lpx_write_lpt which allows writing LP problem
1.1735 + data from an LP object to a text file using the CPLEX LP format
1.1736 + was included in the package and documented.
1.1737 +
1.1738 + * glplpx.h, glplpx3.c
1.1739 + The control parameter LPX_K_LPTORIG that affects the behavior
1.1740 + of the api routine lpx_write_lpt was introduced.
1.1741 +
1.1742 + * glplan6.c
1.1743 + The semantics of the language GLPK/L was changed to allow
1.1744 + selection in case when not all mute letters of a predicate (the
1.1745 + operand that follows the keyword 'where') are presented in a
1.1746 + parameter (the operand that precedes the keyword 'where'), i.e.
1.1747 + to allow writing something like this:
1.1748 + y[j] := sum(i, x[i] where p[i,j]);
1.1749 + The paragraph "Selection" in the langauge description (page 25)
1.1750 + was also correspondingly changed. This change of the language
1.1751 + semantics was undertaken due to a notice provided by Peter Lee
1.1752 + <plee@kinggee.com.au>.
1.1753 +
1.1754 + * sample/hwd.lpm
1.1755 + A nice example of LP model written in GLPK/L and contributed by
1.1756 + Peter Lee <plee@kinggee.com.au> was included in the package.
1.1757 +
1.1758 + * glplpx6b.c
1.1759 + The api routine lpx_interior was modified: a) to compute dual
1.1760 + values for all structural as well as auxiliary variables; b) to
1.1761 + allow specifying non-zero objective coefficients at auxiliary
1.1762 + variables.
1.1763 +
1.1764 + * sample/glpsol.c
1.1765 + Three new command-line options were added to the solver, which
1.1766 + are: --plain, --orig, and --wrlpt.
1.1767 +
1.1768 +Mon Oct 14 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1769 +
1.1770 + * GLPK 3.2.2 has been released.
1.1771 +
1.1772 + * glplpt.h, glplpt.c
1.1773 + A module that reads LP/MIP problem data in CPLEX LP format was
1.1774 + implemented.
1.1775 +
1.1776 + * glplpx8c.c
1.1777 + An api routine lpx_read_lpt that reads LP/MIP problem data in
1.1778 + CPLEX LP format was implemented.
1.1779 +
1.1780 + * sample/glpsol.c, sample/plan.lpt
1.1781 + A new command-line option '--lpt' that allows reading LP/MIP
1.1782 + problem data in CPLEX LP format was added to the solver.
1.1783 +
1.1784 + * doc/refman.latex, doc/refman.dvi, doc/refman.ps
1.1785 + A new edition of the Reference Manual was included.
1.1786 +
1.1787 + * source/*.c
1.1788 + Casting to (unsigned char) was added in some calls to the
1.1789 + classification functions (isalpha, etc.). The bug was fixed due
1.1790 + to report provided by Morten Welinder <terra@diku.dk>.
1.1791 +
1.1792 + * glplpx8a.c
1.1793 + The local routine mps_numb used in the routine lpx_write_mps
1.1794 + was modified to correctly format floating-point numbers that
1.1795 + have two digits in the decimal exponent. The bug was fixed due
1.1796 + to report provided by Vlahos Kiriakos <Kiriakos.Vlahos@gs.com>.
1.1797 +
1.1798 + * glplan.h, glplan1.c, ..., glplan8.c
1.1799 + Several serious bugs were fixed in the language processor due
1.1800 + to reports provided by <NORBERT.PIOTROWSKI@LHSYSTEMS.COM>:
1.1801 + (a) a static search tree used to find sparse array elements was
1.1802 + sometimes overwritten that caused the message 'assertion failed'
1.1803 + to appear; the bug was fixed by creating separate search trees
1.1804 + in parsing routines; (b) a variable declared using the
1.1805 + predicate-controlled variable declaration statement had wrong
1.1806 + order of domain sets, because the variable array was built as
1.1807 + a copy of the predicate array; the bug was fixed by using the
1.1808 + internal routine transpose that coordinates mute letters (and
1.1809 + therefore domain sets) on copying sparse arrays; (c) sometimes
1.1810 + assignment statements like x[#a,#b,#c] := ... was incorrectly
1.1811 + processed; the bug was fixed by including an appropriate check
1.1812 + into the internal routine assign_stmt.
1.1813 +
1.1814 + * glp_simplex.c
1.1815 + An additional check to see if all lower bounds are not greater
1.1816 + than corresponding upper bounds was included in the routine to
1.1817 + prevent wrong results to appear. Such incorrectness sometimes
1.1818 + was not detected, namely, when variables with such bounds were
1.1819 + non-basic and never entered the basis.
1.1820 +
1.1821 + * glpspx1.c
1.1822 + Maximal number of simplex iterations before reinversion was
1.1823 + decreased from 100 to 50. This allowed to improve accuracy and,
1.1824 + that is more important, to reduce the solution time for many
1.1825 + serial lp problems approximately 1.5--2 times.
1.1826 +
1.1827 + * glpspx2.c
1.1828 + A check to see if all elements in the column chosen to enter
1.1829 + the basis are close to zero in the routine spx_prim_chuzr was
1.1830 + temporarily removed because this check gave wrong conclusion in
1.1831 + case when the corresponding non-basic variable had zero column
1.1832 + in the constraint matrix. An analogous check to see if all
1.1833 + elements in the row chosen to leave the basis are close to zero
1.1834 + in the routine spx_dual_chuzc was also temporarily removed on
1.1835 + the same reason. The bug was fixed due to reports provided by
1.1836 + Flavio Keidi Miyazawa <fkm@ic.unicamp.br> and Vlahos Kiriakos
1.1837 + <Kiriakos.Vlahos@gs.com>.
1.1838 +
1.1839 +Mon Aug 12 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1840 +
1.1841 + * GLPK 3.2.1 has been released.
1.1842 +
1.1843 + * glpbcs.h, glpbcs1.c, glpbcs2.c
1.1844 + * glpies.h, glpies1.c, glpies2.c, glpies3.c
1.1845 + A preliminary implementation of the branch-and-cut framework
1.1846 + was included in the package.
1.1847 +
1.1848 + * doc/brcut.txt
1.1849 + The document "GLPK: A Preliminary Implementation of the
1.1850 + Branch-And-Cut Framework" was included in the distribution.
1.1851 +
1.1852 + * sample/tspsol.c
1.1853 + An illustrative program for solving symmetric TSP based on the
1.1854 + branch-and-cut method was included in the package.
1.1855 +
1.1856 + * glpdmp.h, glpdmp.c
1.1857 + A new, re-enterable version of routines for managing dynamic
1.1858 + memory pools was included in the package.
1.1859 +
1.1860 + * glpavl.h, glpavl.c
1.1861 + A new, re-enterable version of routines for managing AVL search
1.1862 + trees was included in the package.
1.1863 +
1.1864 + * glplib.h, glplib2.c
1.1865 + Two new low-level routines ufopen and ufclose were included in
1.1866 + the package.
1.1867 +
1.1868 + * glplpx.h, glplpx7.c
1.1869 + The following new api routines were added: lpx_eval_activity,
1.1870 + lpx_eval_red_cost, lpx_reduce_form, lpx_mixed_gomory.
1.1871 +
1.1872 + * glptsp.h, glptsp.c
1.1873 + A module for reading TSP data using TSPLIB format was included
1.1874 + in the package.
1.1875 +
1.1876 +Mon Jul 15 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1877 +
1.1878 + * GLPK 3.2 has been released.
1.1879 +
1.1880 + * glplpx.h, glplpx1.c, glplpx2.c
1.1881 + The identifier 'class' (used as a member name in the structure
1.1882 + LPX and as an argument name in the routine lpx_set_class) was
1.1883 + changed to 'clss' in order to avoid conflicts with C++ reserved
1.1884 + words.
1.1885 +
1.1886 + * glpk.h, glplpx.h, glplpx1.c, glplpx2.c, glplpx6a.c,
1.1887 + * glplpx6b.c, glplpx6c.c, glplpx7.c, glplpx8.c
1.1888 + The following new api routines were added: lpx_set_obj_name,
1.1889 + lpx_get_obj_name, lpx_get_row_mark, lpx_get_col_mark,
1.1890 + lpx_transform_row, lpx_transform_col, lpx_prim_ratio_test,
1.1891 + lpx_dual_ratio_test, lpx_interior, lpx_get_ips_stat,
1.1892 + lpx_get_ips_row, lpx_get_ips_col, lpx_get_ips_obj, lpx_read_lpm,
1.1893 + lpx_write_mps, lpx_print_ips.
1.1894 +
1.1895 + * glpsol.c
1.1896 + The solver was completely re-programmed using new api routines.
1.1897 +
1.1898 + * lang.latex, lang.dvi, lang.ps
1.1899 + New edition of the document "GLPK: Modeling Language GLPK/L"
1.1900 + was included in the distribution.
1.1901 +
1.1902 + * refman.latex, refman.dvi, refman.ps
1.1903 + New edition of the document "GLPK: Reference Manual" (which
1.1904 + contains descriptions of all new api routines) was included in
1.1905 + the distribution.
1.1906 +
1.1907 + * glpapi.h, glpapi1.c, glpapi2.c, glpapi3.c, glpapi4.c
1.1908 + These files (which contain old api routines) were removed from
1.1909 + the package.
1.1910 +
1.1911 + * glpipm1.c, glpipm2.c
1.1912 + The file glpipm1.c was renamed to glpipm.c. The file glpipm2.c
1.1913 + was used only by old api routines and therefore was removed from
1.1914 + the package.
1.1915 +
1.1916 + * language.texinfo
1.1917 + Old version of the document "GLPK: Modeling Language GLPK/L" was
1.1918 + removed from the distribution.
1.1919 +
1.1920 +Mon May 27 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1921 +
1.1922 + * GLPK 3.1 has been released.
1.1923 +
1.1924 + * glplpx.h, glplpx1.c, glplpx2.c, glplpx3.c, glplpx4.c,
1.1925 + * glplpx5.c, glplpx6.c, glplpx7.c, glplpx8.c
1.1926 + A preliminary implementation of new API routines was completed.
1.1927 +
1.1928 + * refman.latex, refman.dvi, refman.ps
1.1929 + A draft edition of the document "GLPK Reference Manual", which
1.1930 + describes new API routines, was included.
1.1931 +
1.1932 + * glplib3.c
1.1933 + A bug in measuring long time intervals was fixed up.
1.1934 +
1.1935 + * glprsm3.c
1.1936 + This module contains some obsolete routines not longer used and
1.1937 + therefore it was removed from the package (into the subdirectory
1.1938 + 'oldsrc').
1.1939 +
1.1940 + * glprsm.h
1.1941 + Some declarations related to the module 'glprsm3.c' (see above)
1.1942 + were removed.
1.1943 +
1.1944 + * guide.texinfo
1.1945 + The document "GLPK User's Guide" describing old API routines was
1.1946 + removed from the package (into the subdirectory 'oldsrc').
1.1947 +
1.1948 + * newapi.txt
1.1949 + The document "New GLPK API Routines" was removed at all, because
1.1950 + it is superseded by the new reference manual (see above).
1.1951 +
1.1952 +Mon May 13 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1953 +
1.1954 + * GLPK 3.0.8 has been released.
1.1955 +
1.1956 + * glplpx.h, glplpx1.c, glplpx2.c, glplpx3.c, glplpx4.c,
1.1957 + * glplpx5.c, glplpx6.c, glplpx7.c
1.1958 + A preliminary (currently incomplete) implementation of new api
1.1959 + routines was included.
1.1960 +
1.1961 + * sample/newsamp.c
1.1962 + A sample program for the new api routines was included.
1.1963 +
1.1964 + * newapi.txt
1.1965 + A draft of the document "New GLPK API Routines" was included.
1.1966 +
1.1967 + * glpapi2.c, glpapi5.c, glpapi6.c
1.1968 + These modules (which contain the api routines glp_call_rsm1,
1.1969 + glp_simplex1, glp_pivot_in, glp_pivot_out) were removed from the
1.1970 + package (to the subdirectory 'oldsrc') since these routines are
1.1971 + functionally superseded by the new api routines.
1.1972 +
1.1973 + * glpk.h, glpapi2.c, glpapi3.c, glpapi4.c
1.1974 + The api routines glp_simplex2, glp_call_ipm1, glp_call_bbm1 were
1.1975 + renamed to glp_simplex, glp_interior, glp_integer, respectively.
1.1976 +
1.1977 + * sample/glpsol.c
1.1978 + Some command-line options (which got obsolete due to the recent
1.1979 + changes in api) were excluded.
1.1980 +
1.1981 + * doc/guide.texinfo
1.1982 + New edition of the document "GLPK User's Guide" was included in
1.1983 + the distribution to reflect the changes in some api routines.
1.1984 +
1.1985 + * doc/libref.texinfo
1.1986 + This document was removed from the package (to the subdirectory
1.1987 + 'oldsrc') since it describes the library routines, most of which
1.1988 + got obsolete and no longer used.
1.1989 +
1.1990 + * Makefile.in
1.1991 + A minor bug was fixed up due to bug report from Hans Schwengeler
1.1992 + <Hans.Schwengeler@unibas.ch>.
1.1993 +
1.1994 +Mon Apr 22 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.1995 +
1.1996 + * GLPK 3.0.7 has been released.
1.1997 +
1.1998 + * glpduff.h, glpduff.c, glpspx.h, glpspx1.c, glpspx2.c,
1.1999 + * glpapi7.c
1.2000 + These modules were replaced by a new implementation of the
1.2001 + simplex method and therefore they were removed from the package
1.2002 + (however they still can be found in the subdirectory 'oldsrc').
1.2003 +
1.2004 + * glprsm1.c
1.2005 + The routine crash_aa was replaced by a new implementation and
1.2006 + therefore it was removed from the file 'glprsm1.c'.
1.2007 +
1.2008 + * glplpx.h, glplpx.c, glpspx.h, glpspx1.c, glpspx2.c, glpspx3.c,
1.2009 + * glpspx4.c, glpapi7.c
1.2010 + New (currently incomplete) implementation of the simplex method
1.2011 + components was included in the package.
1.2012 +
1.2013 +Thu Mar 28 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2014 +
1.2015 + * GLPK 3.0.6 has been released.
1.2016 +
1.2017 + * glpluf.h, glpluf.c, glpinv.h, glpinv.c
1.2018 + New version of LU-factorization and basis maintenance routines
1.2019 + (based on Forrest-Tomlin updating technique) was implemented.
1.2020 +
1.2021 + * glpeta.h, glpeta.c, glpfhv.h, glpfhv.c, glpgel.h, glpgel.c,
1.2022 + * glppfi.h, glppfi.c, glprfi.h, glprfi.c
1.2023 + These routines implement some other forms of the basis matrix.
1.2024 + Now they became obsolete being functionally superseded by the
1.2025 + new version of basis maintenance routines (see above) and were
1.2026 + removed from the package (however they still can be found in the
1.2027 + subdirectory 'oldsrc').
1.2028 +
1.2029 + * glpbbm.c, glprsm.h, glprsm1.h, glprsm2.h, glpspx.h, glpspx2.c,
1.2030 + * glprsm2.c, glpsol.c
1.2031 + Necessary changes were made in order to use the new version of
1.2032 + basis maintenance routines.
1.2033 +
1.2034 +Tue Jan 29 12:00:00 2002 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2035 +
1.2036 + * GLPK 3.0.5 has been released.
1.2037 + Structure of the package was re-organized in order to simplify
1.2038 + its maintenance.
1.2039 +
1.2040 + * doc/guide.texinfo
1.2041 + New edition of the document "GLPK User's Guide" was included in
1.2042 + the distribution. Now the document includes descriptions of some
1.2043 + additional API routines recently added to the package.
1.2044 +
1.2045 + * doc/newapi.txt
1.2046 + The document "Additional GLPK API Routines" was removed from the
1.2047 + distribution, because the corresponding material was included in
1.2048 + the user's guide (see above).
1.2049 +
1.2050 +Mon Dec 10 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2051 +
1.2052 + * GLPK 3.0.4 has been released.
1.2053 +
1.2054 + * glpspx.h, glpspx1.c, glpspx2.c, glpapi/glp_simplex2.h
1.2055 + A new, more efficient version of the two-phase primal simplex
1.2056 + method was implemented (advanced initial basis, projected
1.2057 + steepest edge, recursive computations of solution components).
1.2058 +
1.2059 + * glpapi/glp_call_bbm1.c
1.2060 + Now LP relaxation can be solved either using rsm1_driver(), or
1.2061 + using glp_simplex2(). The choice is controlled by the parameter
1.2062 + 'meth' (a member of struct bbm1).
1.2063 +
1.2064 + * sample/glpsol.c
1.2065 + The new implementation of the simplex method is now used by
1.2066 + default. The old version is available via --old-sim option.
1.2067 +
1.2068 + * glpmat/gm_scaling.c
1.2069 + Now this routine displays only two lines: an initial "quality"
1.2070 + and a final "quality".
1.2071 +
1.2072 + * glplp/prepro_lp.c
1.2073 + Identifiers 'fmin' and 'fmax' renamed to 'f_min' and 'f_max' in
1.2074 + order to avoid conflict with <math.h>. The bug was fixed due to
1.2075 + report provided by Sami Farin <sfarin@ratol.fi>.
1.2076 +
1.2077 +Wed Oct 03 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2078 +
1.2079 + * GLPK 3.0.3 has been released.
1.2080 +
1.2081 + * glprsm/harris_row.c, glprsm/harris_col.c
1.2082 + The relative tolerance used on the first pass of the two-pass
1.2083 + ratio test was replaced by the absolute tolerance.
1.2084 +
1.2085 + * glprsm/rsm_primal.c, glprsm/rsm_feas.c, glprsm/rsm_dual.c
1.2086 + The absolute tolerance passed to the two-pass ratio test routine
1.2087 + was decaresed (for both primal and dual simplex).
1.2088 +
1.2089 + These changes were made in order to improve numerical stability
1.2090 + of the simplex method.
1.2091 +
1.2092 + * glprsm/glp_call_rsm1.c, glprsm/glp_call_bbm1.c,
1.2093 + * glprsm/glp_simplex1, glprsm/glp_pivoting.c
1.2094 + Default form of the inverse was changed from RFI to AFI.
1.2095 +
1.2096 +Mon Sep 24 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2097 +
1.2098 + * GLPK 3.0.2 has been released.
1.2099 +
1.2100 + * glpfhv.h, glpfhv.c
1.2101 + New version of the basis maintaining routines was implemented.
1.2102 + These routines, which are based on so called FHV-factorization
1.2103 + (a variety of LU-factorization) and Gustavson's data structures,
1.2104 + perform the main operations on the basis matrix faster at the
1.2105 + expense of some worsening numerical accuracy.
1.2106 +
1.2107 + * glprsm.h, glprsm/afi.c
1.2108 + The routines, which implement AFI (Advanced Form of the
1.2109 + Inverse) based on FHV-factorization, were added to the package.
1.2110 + This new form is available via the parameter form = 3 (on API
1.2111 + level) or via the option --afi (in GLPSOL solver).
1.2112 +
1.2113 + * EFI was renamed to PFI
1.2114 + In order to correct terminology the acronym EFI (Elimination
1.2115 + Form of the Inverse) was replaced by PFI (Product Form of the
1.2116 + Inverse) everywhere in the source code and the documentation.
1.2117 +
1.2118 + * glpset/umalloc.c, glpset/ucalloc.c
1.2119 + * glpset/get_atom.c, glpset/get_atomv.c
1.2120 + These memory management routines were changed in order *not* to
1.2121 + clear allocated memory blocks by binary zeros.
1.2122 +
1.2123 +Wed Aug 01 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2124 +
1.2125 + * GLPK 3.0.1 has been released.
1.2126 +
1.2127 + * glpapi/old_api.c, glplp/extract_lp.c, store_lpsol.c
1.2128 + Old API routines were deleted from the package.
1.2129 +
1.2130 + * include/glpk.h, include/glpapi.h, include/glplp.h
1.2131 + Specifications of old API routines and data structures were
1.2132 + removed from the headers.
1.2133 +
1.2134 + * sample/glpsol.c
1.2135 + New version of the stand-alone solver GLPSOL that now uses new
1.2136 + API routines was implemented.
1.2137 +
1.2138 + * glpapi/glp_set_row_fctr.c, glpapi/glp_set_col_fctr.c,
1.2139 + * glpapi/glp_get_row_fctr.c, glpapi/glp_get_col_fctr.c,
1.2140 + * glpapi/glp_scale_prob.c
1.2141 + Scaling routines were added.
1.2142 +
1.2143 + * glpapi/glp_write_mps.c
1.2144 + The routine for writing problem data in MPS format was added.
1.2145 +
1.2146 + * glpapi/glp_simplex1.c
1.2147 + Comprehensive driver to the simplex method was added.
1.2148 +
1.2149 + * glpapi/glp_pivoting.c
1.2150 + The routines glp_pivot_in() and glp_pivot_out() intended for
1.2151 + basis maintaining were added.
1.2152 +
1.2153 + * glprsm/create_rsm.c, glprsm/delete_rsm.c, glprsm/scale_rsm.c,
1.2154 + * glprsm/build_basis.c
1.2155 + Additional low level routines related to the simplex method
1.2156 + were added.
1.2157 +
1.2158 + * glpk.h, glpapi.h, glprsm.h
1.2159 + Additional specifications for new routines and data structures
1.2160 + were added.
1.2161 +
1.2162 + * sample/lpglpk30.c
1.2163 + A non-trivial example was added. It allows using GLPK as a base
1.2164 + LP solver for Concorde, a program for solving Traveling Salesman
1.2165 + Problem (TSP). For details see comments in 'lpglpk30.c'.
1.2166 +
1.2167 + * doc/newapi.txt
1.2168 + The document "Additional GLPK API Routines" that describes some
1.2169 + new API routines was included.
1.2170 +
1.2171 +Thu Jul 19 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2172 +
1.2173 + * GLPK 3.0 has been released.
1.2174 +
1.2175 + Now GLPK is provided with new API, which is intended for using
1.2176 + the package in more complex algorithmic schemes.
1.2177 +
1.2178 + * glpapi/old_api.c
1.2179 + All routines related to old API were gathered in one file named
1.2180 + 'old_api.c'.
1.2181 +
1.2182 + * glpapi/*.c
1.2183 + These routines that implement new API were added to the package.
1.2184 +
1.2185 + * include/glpk.h, include/glpapi.h
1.2186 + Specifications of new API routines and data structures were
1.2187 + added to these headers. Specifications of old API routines and
1.2188 + data structures were locked by #ifdef GLP_OLD_API directive.
1.2189 +
1.2190 + * doc/guide.texinfo
1.2191 + New edition of the document "GLPK User's Guide" that correspond
1.2192 + to new API was included.
1.2193 +
1.2194 +Thu Jun 14 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2195 +
1.2196 + * GLPK 2.4.1 has been released.
1.2197 +
1.2198 + * doc/glpk_ml.texinfo
1.2199 + The new document "Modeling Language GLPK/L" was included.
1.2200 +
1.2201 + * doc/glpk_ug.texinfo
1.2202 + New edition of the document "GLPK User's Guide" was included.
1.2203 +
1.2204 + * doc/language.txt
1.2205 + The preliminary document "GLPK/L Modeling Language: A Brief
1.2206 + description" was removed from the distribution, because it has
1.2207 + been replaced by the new document "Modeling Language GLPK/L".
1.2208 +
1.2209 + * glplang/l_spar.c
1.2210 + The routine comparison() was re-programmed in order to
1.2211 + implement the relation operation as specified in the language
1.2212 + description.
1.2213 +
1.2214 + * glpmip.h, glpmip/*.c
1.2215 + The partition 'glpmip' was renamed to 'glpbbm'.
1.2216 +
1.2217 +Thu May 10 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2218 +
1.2219 + * GLPK 2.4 has been released.
1.2220 +
1.2221 + Now GLPK includes an implementation of a preliminary version of
1.2222 + the GLPK/L modeling language.
1.2223 +
1.2224 + * glplang.h, glplang/*.c
1.2225 + The header 'glplang.h' and a set of routines that implements
1.2226 + the GLPK/L language processor (the partition 'glplang') were
1.2227 + added to the package.
1.2228 +
1.2229 + * doc/language.txt
1.2230 + The document "GLPK/L Modeling Language: A Brief Description
1.2231 + (Supplement to GLPK User's Guide)" in plain text format was
1.2232 + included in the package (see the file 'language.txt' in the
1.2233 + subdirectory 'doc' of the distribution).
1.2234 +
1.2235 + * ex/model1.lpm, ex/model2.lpm
1.2236 + Two examples of model descriptions written in GLPK/L were added
1.2237 + to the package.
1.2238 +
1.2239 + * sample/glpsol.c
1.2240 + This program was modified in order: a) to allow processing
1.2241 + model description written in GLPK/L; b) to allow solving pure
1.2242 + LP problem using the interior point method.
1.2243 +
1.2244 + * sample/glpipm.c
1.2245 + This program was removed from the package, because its function
1.2246 + was passed to the GLPSOL solver.
1.2247 +
1.2248 + * Makefile.in
1.2249 + This file was changed in order to install the GLPSOL solver
1.2250 + executable.
1.2251 +
1.2252 +Mon Apr 09 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2253 +
1.2254 + * GLPK 2.3 has been released.
1.2255 +
1.2256 + * glpmip.h, glpmip/*.c
1.2257 + These routines (that implement the branch-and-bound method) were
1.2258 + re-programmed in order to improve robustness of implementation.
1.2259 + In particular, heuristic routines were carried out from the main
1.2260 + driver routine.
1.2261 +
1.2262 + Additional GLPK API routines were documented.
1.2263 +
1.2264 + New edition of the document "GLPK User's Guide" was included in
1.2265 + the package.
1.2266 +
1.2267 + The preliminary document "Mixed Integer Programming Using GLPK
1.2268 + Version 2.2 (Supplement to GLPK User's Guide)" was removed from
1.2269 + the package, because this material was included in GLPK User's
1.2270 + Guide.
1.2271 +
1.2272 +Thu Mar 15 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2273 +
1.2274 + * GLPK 2.2 has been released.
1.2275 +
1.2276 + Now GLPK includes a tentative implementation of the
1.2277 + branch-and-bound procedure based on the dual simplex method for
1.2278 + mixed integer linear programming (MIP).
1.2279 +
1.2280 + The preliminary document "Mixed Integer Programming Using GLPK
1.2281 + Version 2.2 (Supplement to GLPK User's Guide)" was included into
1.2282 + the package in plain text format (see the file 'mip.txt' in the
1.2283 + subdirectory 'doc' of the distribution).
1.2284 +
1.2285 + * glpmip.h, glpmip/*.c, glpapi/glp_integer.c
1.2286 + These routines (that implement the branch-and-bound method) were
1.2287 + added to the package.
1.2288 +
1.2289 + * sample/glpsol.c
1.2290 + This program was modified in order to allow solving LP and MIP
1.2291 + problems.
1.2292 +
1.2293 + * glprsm/rsm_primal.c, glprsm/rsm_dual.c, glprsm/rsm_feas.c,
1.2294 + * glprsm/rsm1_driver.c
1.2295 + These routines (which are drivers to basic components of the
1.2296 + revised simplex method) were added to the package.
1.2297 +
1.2298 +Mon Feb 19 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2299 +
1.2300 + * GLPK 2.1 has been released.
1.2301 +
1.2302 + * glprsm.h, glprsm/*.c
1.2303 + These routines (that implement components of the revised simplex
1.2304 + method) were re-programmed and documented.
1.2305 +
1.2306 + The document "GLPK Implementation of the Revised Simplex Method"
1.2307 + was included into the package.
1.2308 +
1.2309 +Thu Jan 25 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2310 +
1.2311 + * GLPK 2.0 has been released.
1.2312 +
1.2313 + Now GLPK includes a tentative implementation of the primal-dual
1.2314 + interior point method for large-scale linear programming (for
1.2315 + more details see the file `NEWS' in the distribution). A number
1.2316 + of routines related to the interior point method were added to
1.2317 + the package.
1.2318 +
1.2319 + * insist.c
1.2320 + The routine `insist' and the macro of the same name were
1.2321 + introduced into the package in order to replace the standard
1.2322 + macro `assert'. Some routines require the expression specified
1.2323 + in the `assert' macro to be evaluated, but compiling the package
1.2324 + with NDEBUG option prevents from that. This bug was fixed due to
1.2325 + bug report provided by Peter A. Huegler <phuegler@bsco.com>.
1.2326 +
1.2327 + * Makefile.in
1.2328 + Minor bug was fixed due to a patch provided by Alexandre Oliva
1.2329 + <oliva@lsd.ic.unicamp.br>.
1.2330 +
1.2331 +Wed Jan 10 12:00:00 2001 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2332 +
1.2333 + * GLPK 1.1.2 has been released.
1.2334 +
1.2335 + * umalloc.c, ufree.c, create_pool.c, get_atom.c, get_atomv.c
1.2336 + These routines were changed in order to fix a bug due to
1.2337 + report provided by Andrew Hood <ajhood@fl.net.au>. Because of
1.2338 + this bug data alignment error occured on the Sparc computer.
1.2339 +
1.2340 +Tue Dec 14 12:00:00 2000 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2341 +
1.2342 + * GLPK 1.1.1 has been released.
1.2343 +
1.2344 + Minor bug was fixed in `Makefile.in'.
1.2345 +
1.2346 + GLPK Library Reference was included.
1.2347 +
1.2348 +Mon Nov 27 12:00:00 2000 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2349 +
1.2350 + * GLPK 1.1 has been released.
1.2351 +
1.2352 + Minor changes were made in order to co-ordinate GLPK routines
1.2353 + and their descriptions.
1.2354 +
1.2355 + GLPK User's Guide was included.
1.2356 +
1.2357 +Fri Oct 20 12:00:00 2000 Andrew Makhorin <mao@mai2.rcnet.ru>
1.2358 +
1.2359 + * GLPK 1.0 has been released.