Comparison of OPT-values for instances at https://github.com/Bansal-ORGroup/MCLS-LSPC-Data provided by Kartik Kulkarni & Manish Bansal (2019) in: "Multi-Module Capacitated Lot-Sizing Problem and its Generalizations with Two-Echelons and Piecewise Concave Production Costs" https://optimization-online.org/2019/07/7294/ and OPT-values calculated with https://www.lutanho.net/plt/lotsize_mcm.html (MCLS instances) and calculated with https://www.lutanho.net/plt/lotsize.html (LSPC instances) (all costs multiplied by 100 to get integer values as input for the solvers) ------------------------------------------------------------- page 33/55 Table 2: instance data of files in MCLS_Instances/2 Machines/ reformatted with https://www.lutanho.net/plt/lotsizing/mcm/instance_generator_MCLS_2Machines.txt and solved with https://www.lutanho.net/plt/lotsize_mcm.html OPT_KK&MB OPT_LT CalcTime_LT_Solver T=40, (C1, C2)=(670, 1280): 1: 105776.2 =10577615 3+684+0 ms okay 2: 102923.9 =10292394 1+676+1 ms okay 3: 103685.4 <10687157 1+1163+0 ms different! file MCLS_Instances/2 Machines/40x1x3.txt 4: 109564.4 =10956437 1+1046+0 ms okay 5: 102607.9 =10260785 2+1489+0 ms okay T=40, (C1, C2)=(850, 1590): 1: 84732.38 =8473238 1+522+0 ms okay 2: 84731.35 =8473135 1+632+0 ms okay 3: 88749.14 =8874914 1+385+0 ms okay 4: 84815.14 =8481514 1+524+0 ms okay 5: 84755.59 =8475559 0+431+1 ms okay T=60, (C1, C2)=(960, 1970): 1: 109912.4 =10991240 1+674+0 ms okay 2: 112890.9 =11289090 2+1302+0 ms okay 3: 113122.8 >11301895 2+685+0 ms different! file MCLS_Instances/2 Machines/60x1x3.txt (Gusek:113018.9->okay, see below) 4: 113003.9 =11300394 2+704+0 ms okay 5: 113014.8 =11301478 1+656+0 ms okay T=60, (C1, C2)=(870, 1590): 1: 122124.1 =12212410 3+2523+0 ms okay 2: 124339.7 >12059881 2+972+0 ms different! file MCLS_Instances/2 Machines/60x2x2.txt (Gurobi:CNS) 3: 125196.7 =12519671 2+1040+0 ms okay 4: 125655.1 =12565514 2+1510+1 ms okay 5: 119757.9 <12523282 3+2115+0 ms different! file MCLS_Instances/2 Machines/60x2x5.txt (Gurobi:CNS) T=80, (C1, C2)=(960, 1970): 1: 153922.4 =15392238 3+2344+0 ms okay 2: 151669.2 >14952779 2+2125+0 ms different! file MCLS_Instances/2 Machines/80x1x2.txt (Gurobi:CNS) 3: 153344.7 >15036283 2+1547+0 ms different! file MCLS_Instances/2 Machines/80x1x3.txt (Gurobi:CNS) 4: 157264.5 >15040454 4+3839+1 ms different! file MCLS_Instances/2 Machines/80x1x4.txt (Gurobi:CNS) 5: 152732.9 >14909798 4+1540+0 ms different! file MCLS_Instances/2 Machines/80x1x5.txt (Gurobi:CNS) T=80, (C1, C2)=(1310, 2570): 1: 119716.1 =11971612 5+1488+0 ms okay 2: 119763.4 =11976335 3+976+0 ms okay 3: 120956 =12095597 2+1334+1 ms okay 4: 120552 =12055199 2+924+0 ms okay 5: 120038.4 =12003836 1+885+0 ms okay Instances for which the OPT-value of my solver is smaller than the provided OPT-value can be verified by using a MIP-solver - solving and adding sub-problem-solutions between zero-inventory periods: Instance; zero-inventory periods; cumulated costs until zero-inventory periods; 40x1x3.txt: not possible to verify this way, OPT-value of my solver is larger 60x1x3.txt: (10, 14,) 18, 37, 60; (1867802, 2597794,) 3309604, 6958420, 11301895; Gusek-OPT:33096+36488.2+43434.7=113018.9 60x2x2.txt: 60; 12059881; (never goes down to inventory=0 inbetween) 60x2x5.txt: not possible to verify this way, OPT-value of my solver is larger 80x1x2.txt: 29, 48, 66, 72, 80; 5528608, 9157503, 12437559, 13539554, 14952779; 80x1x3.txt: 80; 15036283; (never goes down to inventory=0 inbetween) 80x1x4.txt: 31, 60, 80; 5906646, 11349673, 15040454; 80x1x5.txt: 29, 80; 5474579, 14909798; ------------------------------------------------------------- page 34/55 Table 3: instance data of files in MCLS_Instances/3 Machines/ reformatted with https://www.lutanho.net/plt/lotsizing/mcm/instance_generator_MCLS_3Machines.txt and solved with https://www.lutanho.net/plt/lotsize_mcm.html OPT_KK&MB OPT_LT CalcTime_LT_Solver T=40, (C1, C2, C3)=(970, 1950,2810): 1: 75340.11 >7531346 4+594+0 ms different! file MCLS_Instances/3 Machines/40x1x1.txt (Gurobi:CNS)(Gusek:39879.6+18344.5+17089.3=75313.4->okay) 2: 75504.95 =7550495 3+706+0 ms okay 3: 75460.09 =7546009 2+392+0 ms okay 4: 76050.1 =7605010 2+846+0 ms okay 5: 76032.55 =7603255 1+845+0 ms okay T=40, (C1, C2, C3)=(790, 1650, 2410): 1: 86187.6 <8756976 1+1939+0 ms different! file MCLS_Instances/3 Machines/40x2x1.txt (Gurobi:CNS) 2: 85017.72 =8501772 1+801+0 ms okay 3: 87299.93 =8729993 1+1225+0 ms okay 4: 87738.2 =8773820 2+574+0 ms okay 5: 87134.51 =8713451 2+817+0 ms okay T=40, (C1, C2, C3)=(670, 1280,1970): (all very different, probably wrong data) 1: 94971.26 <10271460 1+2757+0 ms different! file MCLS_Instances/3 Machines/40x3x1.txt 2: 92043.93 <10283971 1+4443+0 ms different! file MCLS_Instances/3 Machines/40x3x2.txt (Gurobi:CNS) 3: 92786.66 <10281324 1+5988+0 ms different! file MCLS_Instances/3 Machines/40x3x3.txt 4: 89009.17 <10336476 1+4926+0 ms different! file MCLS_Instances/3 Machines/40x3x4.txt (Gurobi:CNS) 5: 91492.86 <10325598 1+4525+0 ms different! file MCLS_Instances/3 Machines/40x3x5.txt T=40, (C1, C2, C3)=(860, 1650, 2590): 1: 84201.9 >8376467 1+1971+0 ms different! file MCLS_Instances/3 Machines/40x4x1.txt (Gurobi:CNS) 2: 85163.32 =8516332 2+1126+0 ms okay 3: 84273.62 >8423787 0+1248+0 ms different! file MCLS_Instances/3 Machines/40x4x3.txt 4: 81205.9 >8002439 0+2586+0 ms different! file MCLS_Instances/3 Machines/40x4x4.txt 5: 81779.04 =8177904 1+1361+0 ms okay ------------------------------------------------------ page 37/55 Table 6: instance data of files in LSPC_Instances/2-breakpoints/ reformatted with https://www.lutanho.net/plt/lotsizing/mcm/instance_generator_LSPC.txt and solved with https://www.lutanho.net/plt/lotsize.html OPT_KK&MB OPT_LT CalcTime_LT_Solver T=50, (b1, b2)=(800, 1600): 1: 103801.9 =10308188 718 ms okay (80/08 digits flipped!?) 2: 101954.3 =10195434 348 ms okay 3: 100903.2 =10090319 371 ms okay 4: 101574 =10157396 667 ms okay 5: 101762.9 =10176286 401 ms okay T=50, (b1, b2)=(1310, 2570): 1: 76681.88 =7668188 334 ms okay 2: 76464.41 =7646441 432 ms okay 3: 76101.56 =7610156 192 ms okay 4: 77388.97 =7738897 254 ms okay 5: 77948.03 =7794803 363 ms okay T=75, (b1, b2)=(1310, 2570): (all very different, probably wrong data, 76000*(75/50)=114000 should be expected) 1: 101181.6 <11342388 685 ms different! file LSPC_Instances/2-breakpoints/75x1x1.txt 2: 103614.1 <11487382 1334 ms different! file LSPC_Instances/2-breakpoints/75x1x2.txt 3: 102503.5 <11273738 724 ms different! file LSPC_Instances/2-breakpoints/75x1x3.txt 4: 103992.1 <11243579 806 ms different! file LSPC_Instances/2-breakpoints/75x1x4.txt 5: 102477.8 <11642474 823 ms different! file LSPC_Instances/2-breakpoints/75x1x5.txt T=75, (b1, b2)=(1790, 3470): 1: 94168.54 =9416854 627 ms okay 2: 93047.57 =9304757 305 ms okay 3: 94013.08 =9401308 491 ms okay 4: 93807.6 <9410223 500 ms different! file LSPC_Instances/2-breakpoints/75x2x4.txt 5: 93491.1 >9285788 515 ms different! file LSPC_Instances/2-breakpoints/75x2x5.txt