MDDVRPSRCV1_Test_Instance
Description
The proposed test instance is aimed towards VRP problem that addressed in general the multi depots, homogeneous/ heterogeneous vehicles, as well as a set of road capacity for different edges that may or may not sustained a certain level of damages. We apply this dataset for the application of routing under disaster scenario that speaks for the damaged and limited capacity of the edges/ roads in the road network. The instance problem consists of road network mapped in Euclidean distance where the road capacity of edges, depots, shelters, and connecting nodes coordinates as well as demand of shelters are fixed. Meanwhile the "damage" file addressed the damage value of edges within the network due to an event triggered from a single point. This proposed instance is made possible through the author: Anuar, Wadi Khalid, and his PhD research supervisor: Lai Soon, Lee, as well as the contributing instituition: Universiti Utara Malaysia (affiliated to author) & Universiti Putra Malaysia (affiliated to research instituition)
Files
Steps to reproduce
These instructions is also available in the Readme file above. This instances can be downloaded through the zip files: 1) LOAD_INSTANCE NETWORK.zip Each instance is an excel file containing: 1) number of depots, connecting nodes and shelters + (optional fix vehicle number under "M") 2) the coordinate of each node (depot, shelter, connecting nodes) 3) the road capacity of edges in the network 4) the demands for each shelters 5) road network map (figure) This "damage" data associated with the instances can be downloaded through the zip files: 1) LOAD_DAMAGE NETWORK.zip Each damage file is an excel file containing: 1) edges with damage value 2) coordinate of disaster epicenter (e.g. earthquake) A python file served as an example is provided to extract the instance and damage data: 1) LOAD_INTS_DAMG_V3.py The output of these python code is: 1) cost or distance matrix 2) time travel matrix 3) road capacity matrix 4) number of depots, connecting nodes and shelters + (optional fix vehicle number under "M") 5) the coordinate of each node (depot, shelter, connecting nodes) 6) the road capacity of edges in the network 7) the demands for each shelters 8) set of edges 9) edges-> damage value dictionary 10) set of connecting nodes 11) set of shelter nodes 12) set of depot nodes 13) set of vehicles (may be ignored from excel and manually configured) 14) manual set of homogeneous vehicle capacity (may be adapted to heterogeneous with different capacity) the cost matrix is proportional to the distance calculated based on Euclidean distance the time matrix is computed based on the assumption that vehicle moves under constant velocity 90km/h and delay due to the damage value of each edges. REQUIREMENTS 1) Python 2.7. The python code to load the test instance and damage file into python is tested with python 2.7. It is possible to convert the code to python 3. some syntax changes is advised 2) Microsoft excel 3) zip file extraction tools INSTRUCTIONS 1) Download: (a) LOAD_INSTANCE NETWORK.zip (b) LOAD_DAMAGE NETWORK.zip (c) LOAD_INTS_DAMG_V3.py 2) unpack (a) in a different folder. note the path to folder where (a) is unpacked 3) unpack (b) in a different folder. note the path to folder where (b) is unpacked 4) open LOAD_INTS_DAMG_V3.py with python 2.7 5) copy the path (a) in the python code to load instance files at: os.path.abspath( r'copy path to folder here. replace \ with \\' ) and os.path.join( r'copy path to folder here. replace \ with \\',PickInstance ) 6) copy the path (b) in the python code to load "damage" files os.path.abspath( r'copy path to folder here. replace \ with \\' ) and os.path.join( r'copy path to folder here. replace \ with \\',PickDamage ) 7) save the python file 8) run the python file