GitHub - matteojug/Online-Facility-Location-with-Multiple-Advice

Branches Tags

Name		Name	Last commit message	Last commit date
Latest commit History 1 Commit
data		data
mpi-sw		mpi-sw
README		README
defs.cpp		defs.cpp
fotakis.cpp		fotakis.cpp
makefile		makefile
meyerson.cpp		meyerson.cpp
mpi.cpp		mpi.cpp
multiadvice.cpp		multiadvice.cpp
partition.cpp		partition.cpp
run_all_day.py		run_all_day.py
run_all_synth.py		run_all_synth.py
solve_offline.cpp		solve_offline.cpp
solve_online.cpp		solve_online.cpp
solve_synth.cpp		solve_synth.cpp

Repository files navigation

Tested on Ubuntu 20.04.2 LTS, gcc version 9.3.0, Python 3.8.5

To setup the datasets
 - Download raw data (data/raw/setup.sh)
 - Download uber dataset from https://www.kaggle.com/fivethirtyeight/uber-pickups-in-new-york-city) and extract in data/raw

To generate the instances:
 - Run data/generate_instances.sh

To compile the executable:
 - Run make all

To run the experiments:
 - python3 run_all_synth.py -offline -online -m PROC_NUM
 - python3 run_all_day.py -offline -m PROC_NUM
 - python3 run_all_day.py -online -m PROC_NUM
The run logs are saved in logs/, with each line using the format:
algo_name   solution.cost   solution.facility_cost  solution.service_cost   solution.opened_facilities  solution.elapsed_ms