All code for this project is written in Julia v1.8. The language can be downloaded here.
Before running any benchmarks, ensure that the package is instantiated.
julia
using Pkg
Pkg.activate("examples")
Pkg.precompile()
The following models are available: crpyt4, drone-4-1, drone-4-2, gridavoid4, gridavoid10, gridavoid20, nrp8, refuel06, refuel08, refuel20, rocks12
To run all the benchmarks, run the bash script
./benchmark.sh
You can also run each model benchmark separately. For example, to run nrp8, run
julia examples/run_nrp8.jl
The results logs for each benchmark can be found in the original_results folder.
To reproduce results for SARSOP, run benchmark_sarsop.sh.
To reproduce results for STORM, PAYNT, SAYNT, and Overapp use the toolbox artifact from the SAYNT paper. The experiment.py config is available in /original_results/config/experiments.py.
To reproduce the results for PRISM, use the PRISM Model Checker.