LTSRP

Published: 16-10-2020| Version 1 | DOI: 10.17632/bhnx77wx94.1
Contributor:
Nail Tahirov

Description

This dataset is used for the working paper “Routing automated lane-guided transport vehicles in a warehouse handling returns” (LTSRP) authored by Emde, Tahirov, Gendreau and Glock. We generate three different instance sets namely, S, M, and L that correspond proportionally to the real industry case. Instance set S contains small instances with 9 stations. Set M contains medium size instances with 60 stations. These instances correspond most closely to the real-world case. Finally, with 150 stations set L is the largest instances set. The number of depots is 3 in all cases and each set contains 20 instances. There are 60 instances in total. We use two-dimensional coordinates (x_i,y_i) for each station. For set S and L the value of the coordinates are randomly drawn numbers from the interval (1,15) and (1,100), respectively. Depot i=1,…m, is placed in location (2.5+(i-1)∙ 5, 0) (S) and (16.67+(i-1)∙ 33.33, 0) (L), respectively. For the real-world M instances, we reuse the data from our industry partner, randomly fudging the numbers such that they are proportionally correct. Note that, the first 3 nodes (coordinates) of each instance of M set represent depots. Distances between stations (and depots) are then measured via the Euclidean metric. Each text file is structured as follows: NAME: <S_i, M_i, L_i, i=1…20, describes the name of instance of each instance set > DIMENSION: <integer number describes the dimension of the instance> EDGE_WEIGHT_TYPE: <EUC_2D, distance between stations and depots are measured via Euclidean metric> NODE_COORD_SECTION: <contains ID of each station (and depot) and their coordinates (x,y)> EOF: <end of file>

Files