← Back to Data Sets

Supplementary Webpage

A Three-Layer Adaptive Large Neighborhood Search for Scenario-Based Integrated Facility Location and Routing under Service-Time Uncertainty

Soumen Atta
Faculty of Information Technology, University of Jyväskylä, P.O. Box 35 (Agora), Jyväskylä, 40014, Finland
Email: soumen.s.atta@jyu.fi

The table below provides results for all 40 instances. Instance names give the customer count followed by the replicate number: for example, 10_1 is the first 10-customer instance. ALNS means adaptive large neighborhood search and denotes the proposed matheuristic; subscripts G and H identify Gurobi and ALNS, respectively.

Objectives. L is probability-weighted total customer lateness over the 20 stored service-duration scenarios, measured in the instance time units; C is facility-opening plus routing cost, measured in the instance cost units. LG and CG are recomputed from Gurobi’s returned routes. Best LH and Best CH are the two components of the same best ALNS solution among 30 runs with different random seeds, selected by minimizing L first and then C. Zero lateness means no late customer in any of the 20 scenarios.

Optimality gaps. Gap L (%) and Gap C (%) are Gurobi’s relative incumbent-to-bound optimality gaps for the lateness-minimization and cost-minimization phases, respectively; % means percent. These gaps refer to the raw phase objectives, not the route-recomputed objectives or an ALNS-to-Gurobi cost difference. “—” denotes an unavailable cost-phase gap because the cost-minimization phase was not attempted.

Timing. tG is Gurobi’s recorded solve-procedure elapsed time, including phase transitions but excluding initial model construction and queueing. Mean tH is the average elapsed time of one ALNS run across all 30 seeds. 30-run effort is the sum of those 30 run times, not a single-run time; ALNS times exclude batch setup and inter-job waiting. All reported computing times are in seconds, distinct from the instance time units used for L.

Status. OPT means both Gurobi phases were optimal within the configured relative gap tolerance of 0.01% and numerical tolerances. ML1 means a memory limit ended the lateness phase before optimality was certified; the cost phase was not attempted. ML2 means zero lateness was certified, but a memory limit ended the cost phase before optimality was certified. Neither ML1 nor ML2 certifies minimum cost.

Links. The instance name opens its input file. Schedule opens the selected ALNS solution’s graphical route schedule using nominal service durations. Solution opens that same solution in JSON (JavaScript Object Notation), a machine-readable format.

InstanceGurobiALNS
StatusLGCGGap L (%)Gap C (%)tGBest LHBest CHMean tH30-run effortScheduleSolution
10_1 OPT 0.0002,907.233 0.0000.000 4.7 0.0002,940.556 0.380 11.40 View JSON
10_2 OPT 0.0002,955.954 0.0000.000 2.2 0.0003,018.735 0.589 17.67 View JSON
10_3 OPT 0.0002,693.858 0.0000.000 3.8 0.0002,709.377 0.433 12.98 View JSON
10_4 OPT 0.0002,574.647 0.0000.000 3.3 0.0002,574.647 0.414 12.43 View JSON
10_5 OPT 0.0003,101.219 0.0000.000 4.1 0.0003,101.219 0.474 14.23 View JSON
10_6 OPT 0.0002,632.109 0.0000.000 4.2 0.0002,670.297 0.461 13.82 View JSON
10_7 OPT 0.0002,703.702 0.0000.000 1.2 0.0002,712.127 0.470 14.10 View JSON
10_8 OPT 0.0002,862.472 0.0000.000 2.0 0.0002,926.386 0.461 13.82 View JSON
10_9 OPT 0.0003,117.285 0.0000.000 2.2 0.0003,117.285 0.410 12.31 View JSON
10_10 OPT 0.0002,954.457 0.0000.005 5.4 0.0002,954.457 0.598 17.94 View JSON
25_1 OPT 0.0005,331.487 0.0000.000 772.2 0.0005,416.735 2.499 74.97 View JSON
25_2 OPT 0.0004,775.201 0.0000.000 932.9 0.0004,937.285 1.292 38.76 View JSON
25_3 OPT 0.0005,243.730 0.0000.000 420.6 0.0005,308.573 2.160 64.80 View JSON
25_4 OPT 0.0005,243.398 0.0000.000 4,111.5 0.0005,436.377 1.459 43.78 View JSON
25_5 OPT 0.0004,905.327 0.0000.000 512.7 0.0005,045.855 1.723 51.68 View JSON
25_6 OPT 0.0005,166.901 0.0000.000 1,165.3 0.0005,221.563 1.480 44.39 View JSON
25_7 OPT 0.0004,859.575 0.0000.000 238.8 0.0004,914.669 2.222 66.67 View JSON
25_8 OPT 0.0004,681.295 0.0000.000 916.4 0.0004,732.969 0.968 29.03 View JSON
25_9 OPT 0.0004,956.264 0.0000.000 633.5 0.0004,956.264 1.415 42.46 View JSON
25_10 OPT 0.0004,853.626 0.0000.000 211.0 0.0004,912.469 0.981 29.42 View JSON
50_1 ML2 0.0009,947.289 0.0002.555 13,115.4 0.00010,091.653 2.200 66.00 View JSON
50_2 ML2 0.0009,964.657 0.0002.322 11,313.2 0.00010,050.750 3.461 103.84 View JSON
50_3 ML2 0.00010,577.165 0.0002.032 15,332.8 0.00010,557.167 3.448 103.45 View JSON
50_4 ML2 0.00010,366.258 0.0003.045 18,014.2 0.00010,419.249 2.949 88.46 View JSON
50_5 ML2 0.00010,646.050 0.0002.423 16,762.0 0.00010,652.325 2.659 79.77 View JSON
50_6 ML1 1,463.43013,368.393 100.000— 2,924.3 0.00010,726.385 2.207 66.22 View JSON
50_7 ML1 52.51713,279.240 100.000— 1,890.6 0.00011,216.759 2.641 79.24 View JSON
50_8 ML2 0.0009,450.058 0.0002.135 7,711.3 0.0009,619.421 3.134 94.03 View JSON
50_9 ML2 0.00010,301.920 0.0004.287 6,895.6 0.00010,359.302 3.078 92.33 View JSON
50_10 ML2 0.00010,681.316 0.0003.203 6,950.2 0.00010,652.038 3.099 92.98 View JSON
75_1 ML1 6,268.28419,698.225 100.000— 4,697.6 0.00014,735.949 4.840 145.19 View JSON
75_2 ML1 757.02419,496.320 100.000— 5,831.0 0.00015,289.053 3.919 117.58 View JSON
75_3 ML1 1,598.88019,970.054 100.000— 8,621.1 0.00015,330.739 4.102 123.07 View JSON
75_4 ML1 8,452.64120,091.449 100.000— 13,187.3 0.00015,808.365 3.869 116.08 View JSON
75_5 ML1 4,415.64720,239.317 100.000— 7,370.2 0.00016,029.831 4.322 129.65 View JSON
75_6 ML1 29,302.33118,919.473 100.000— 2,574.3 0.00015,128.174 4.643 139.29 View JSON
75_7 ML1 5,306.32619,852.647 100.000— 3,064.2 0.00016,346.656 4.294 128.81 View JSON
75_8 ML1 263.53219,632.691 100.000— 13,503.1 0.00015,895.765 4.706 141.17 View JSON
75_9 ML1 110.63619,975.249 100.000— 13,817.3 0.00016,298.072 3.884 116.51 View JSON
75_10 ML1 177.94618,510.584 100.000— 5,924.4 0.00014,841.792 5.250 157.50 View JSON

This page was created on January 23, 2026.

This page is maintained by Dr. Soumen Atta, Ph.D.