MIPLIB 2010 version NEXT ========================= instances: - germanrr and ns1663818 solved in 3000 and 834 seconds, respectively, with Gurobi 7.5.1 (8 threads), moved from hard to easy - gmut-75-50 and gmut-77-40 solved with ParaXpress in a 12288 core supercomputer run on HLRN III, moved from open to hard - ger50_17_trans solved by ParaXpress in 30 minutes, moved from open to hard - satellites3-40-fs solved in less than one hour by Gurobi 7.5.1 and CPLEX 12.7.1, moved from hard to easy - neos-937815 solved in less than one hour by Gurobi 7.5.2, moved from open to easy - ns1778858 solved by GCG 3.0 in 4176 seconds, moved to hard - atm20-100 solved by SAS/OR 15.1 in 461 seconds, moved from hard to easy - ns1778858 solved by SAS/OR 15.1 in 746 seconds, moved from hard to easy - ns903616 solved by SAS/OR 15.1 in 1651 seconds, moved from open to easy - first feasible solution for f2000 found by memcomputing MIPLIB 2010 version 1.1.4 ========================= features: - added support for MATLAB (R2017a) MIPLIB 2010 version 1.1.3 ========================= features: - fixed issue in Xpress script instances: - rmine10 was solved by ParaSCIP on HLRN III and Titan with 48 restarted jobs - triptim3 was solved Gurobi 6.5 (12 threads) in about 13 days, moved from open to hard - sct32 was solved by Gurobi 7.0 (32 threads) in about 65 hours, moved from open to hard - usAbbrv-8-25_70 was solved by Gurobi 7.0 (32 threads) in about 29 hours, moved from open to hard - ns1696083 solved by Gurobi 7.0 (12 threads) in about half an hour, moved from hard to easy - seymour solved by Gurobi 7.0 (12 threads) in about 50 minutes, moved from hard to easy - rail02 solved by Xpress 8.1 (8 threads) in about 30 minutes, moved from hard to easy - pigeon12 and pigeon13 solved by Xpress 8.1 (12 threads) in about one second, moved from hard to easy - pigeon19 solved by Xpress 8.1 (12 threads) in about two seconds, moved from open to easy MIPLIB 2010 version 1.1.2 ========================= scripts: - fixed issue in Gurobi script - script now prints information about the CPU used instances: - rococoC11-011100 was solved by CPLEX 12.6.2 (8 threads) in 3346 seconds, moved from hard to easy - reblock420 was solved by Xpress 7.9 (8 threads) in 2967 seconds, moved from hard to easy - lotsize and opm2-z10-s2 solved by Gurobi 6.5 (8 threads) in 35 and 56 minutes, respectively, moved from hard to easy - opm2-z11-s8 solved by Gurobi 6.5 (12 threads) in 37 minutes, moved from hard to easy - triptim3 was solved by ParaSCIP on a supercomputer with 864 cores in 9.5 hours, moved from open to hard - triptim3 was solved Gurobi 6.5 (8 threads) in 40 minutes, moved from hard to easy MIPLIB 2010 version 1.1.1 ========================= scripts: - corrected solution file for benchmark set instances: - a1c1s1 was solved by CPLEX 12.6.2 (8 threads) in 608 seconds, moved from hard to easy MIPLIB 2010 version 1.1.0 ========================= features: - support automatic permutation of instances and test on sets of permuted instances - fixed issue in Xpress script instances: - eilA101-2 and rmatr200-p5 were solved by Xpress 28.01 (12 threads) in 3550 and 1600 seconds, respectively, moved from hard to easy MIPLIB 2010 version 1.0.5 ========================= features: - fixed issue in Xpress script instances: - ns2122603 solved by CPLEX 12.6, moved from open to hard - neos-1140050 was shown to be infeasible by CPLEX 12.6, moved from open to hard and to the infeasible set - neos-984165 solved by Gurobi 6.0, moved from open to hard - set3-10 and set3-20 moved solved by SCIP 3.1.1 parallelized by UG 0.7.5, moved from open to hard MIPLIB 2010 version 1.0.4 =============== features: - added information about feasibility of instances and lists of feasible and infeasible instances and list of instances with unknown feasibility - added optimal solutions for most of the easy and hard instances - fixed bug in run_gurobi.sh (for Gurobi 5.6.3 and higher) instances: - buildingenergy was solved using a search procedure built on top of Cplex 12.2 and by Gurobi 4.6.1 (12 threads), moved from open to easy - toll-like was solved with Gurobi 4.6 (8 threads) after a variable transformation to reduce symmetry, moved from open to hard - first feasible solutions found for in and ns1663818 - infeasibility of lrsa120 proven with Gurobi 4.6.1 - r80x800 was solved with Gurobi 4.6.1 (12 threads) in about 79 hours, moved from open to hard - blp-ic97 was solved with Gurobi 4.6.1 (12 threads) in about 31 minutes and moved from hard to easy - neos-631710 has been solved with Gurobi 4.6.1 (12 threads) in about 21 minutes and moved from open to easy - g200x740i, p100x588b, and vpphard2 solved with CPLEX 12.4, ns1696083 and uc-case3 solved with Gurobi 5.0, moved from open to hard - maxgasflow and uct-subprob solved within one hour with Gurobi 5.0, moved from hard to easy - berlin_5_8_0 solved in 9 hours with CPLEX 12.3 - hanoi5 solved due to the 2012 pseudo-Boolean competition - dg012142 solved by ParaSCIP in 43 hours - set3-15 solved in 64.5 hours with CPLEX 12.4 - dc1c solved in 1100 hours by ParaSCIP - germany50-DBM solved by ParaSCIP in 590 hours, moved from open to hard - bnatt moved from hard to easy - rail03 moved from open to hard, solved by CPLEX 12.4 in 170 hours - shs1023 solved by CPLEX 12.5 in 114.5 hours, moved from open to hard - bg512142 solved by ParaCPLEX in about 490 hours, moved from open to hard - infeasibility of neos-807456 proven, moved from open to hard - queens-30 moved from open to hard - go19 and toll-like solved by Gurobi 5.5 in less than 1 hour, moved to easy - tw-myciel4, ns1111636, and germanrr solved by CPLEX 12.5.1, moved from open to hard - janos-us-DDM solved by AIMMS 3.14 FR1 + CPLEX/Gurobi, moved from open to hard - n3-3 solved by CPLEX 12.6, moved from open to hard - instance in solved by Gurobi development version, moved from open to hard - n9-3 and leo2 solved with CPLEX 12.6 in less than one hour, moved from hard to easy - dolom1 has been solved using ParaSCIP with SCIP 3.0.1 linked to CPLEX 12.5, moved from open to hard - vpphard2 has been solved using Gurobi 5.6.2 (12 threads) in about 2 minutes, moved from hard to easy - berlin_5_8_0 has been solved using CPLEX 12.6. (48 threads) in about 55 minutes, moved from hard to easy - ns1663818 has been solved using SCIP 3.1.0 (with SoPlex as the underlying LP solver) in about 10 days, moved from open to hard - corrected optimal values for momentum1 and ran14x18-disj-8 MIPLIB 2010 version 1.0.3 =============== features: - scripts for lp_solve and GLPK added instances: - added INTEND marker to instances where it was missing (circ10-3, ex1010-pi, f2000, hanoi5, pb-simp-nonunif, sct1, sct32, sct5, and wnq-n100-mw99-14) - runs with Gurobi 4.6.1 and 12 threads changed some solution statuses: b2c1s1, maxgasflow, opm2-z10-s2, opm2-z11-s8, opm2-z12-s14, opm2-z12-s7, rmatr200-p10, satellites3-40-fs, satellites3-40, wnq-n100-mw99-14 moved from open to hard, transportmoment moved from open to easy MIPLIB 2010 version 1.0.2 =============== features: - support Gurobi version 4.5.0 and higher instances: - ns2122603 turned out to be feasible, removed from infeasible test set - runs with CPLEX 12.3 and 12 threads changed some solution statuses: blp-ic97 moved from open to hard, 50v-10, atlanta-ip, co-100, ic97_potential, neos-1620770, nobel-eu-DBE, npmv07, ns1685374, ofi, rococoB10-011000, and satellites2-60 moved from hard to easy - runs with Gurobi 4.5.1 and 12 threads changed some solution statuses: d10200, rococoC11-011100 and seymour-disj-10 moved from open to hard, m100n500k4r1, momentum1, n3div36, ns1644855, nu60-pr9, and sp97ar moved from hard to easy