READ ME File For Dataset: Test Instances for Queue constrained packing: a vehicle ferry case study

Dataset DOI: 10.5258/SOTON/D1130

ReadMe Author: Christine Currie, University of Southampton

This dataset supports the publication:
AUTHORS: Christopher Bayliss, Christine S.M. Currie, Julia A. Bennell, Antonio Martinez-Sykora,
TITLE: Queue-constrained packing: A vehicle ferry case study,
JOURNAL: European Journal of Operational Research,
PAPER DOI: https://doi.org/10.1016/j.ejor.2020.07.027

Parse files by splitting each line on the token "=", then split the second part by the token "," to obtain the individual parameter values.

The directory "testInstances_QueueConstrainedPacking" contains ferry dimensions, vehicle mixes, vehicle dimensions, vehicle type quantities, yard queue width information for each of the 100 test instance in the three classes of instances. The input values and their names are specified in each file, that is, they are self contained. The data is organised with a single file for each individual instance.
yard_queue_length is a function of the number of yard queues, which is an option input parameter (1<=), given by the equation: yard_queue_length=(3*total_area_of_vehicles)/(yard_queue_width*number_of_queues).

The directory "bestMetaheuristicParameters" contains full details of the metaheuristic parameters that worked best for each individual test instance. This directory contains two files, one corresponding to the metaheuristic that utilised the SOPE packing encoder, and one corresponding to the metaheuristic that utilised the GPE packing encoder. Within each file there is a line for for each individual test instance, information that is specified in the file to ensure that the file is self contained. Each line specifies class number, instance number, t0 (simiulated annealing initial temperture parameter), psi_{1-5} (the probabilities that each of the 5 local search neighbourhoods of the packing iteration local search are selected in each simulated annealing iteration), useTriangularDistributionMutations (a binary input which specifies whether a triangule distribution, and not a uniform distribution) is used to generate smaller mutations with a higher probaility in packing iteration local search neighbourhoods, nLanes (the number of yard queues used whilst identifying the best metaheuristic parameters for each instance).


Date of data collection: 2019

Information about geographic location of data collection: University of Southampton, U.K.

Date that the file was created: April 2020
Date that the file was edited to add research output doi: March 2022