swMATH ID: 38051
Software Authors: Robert J. Moss
Description: POMDPStressTesting.jl is a package that uses reinforcement learning and stochastic optimization to find likely failures in black-box systems through a technique called adaptive stress testing. Adaptive stress testing (AST) has been used to find failures in safety-critical systems such as aircraft collision avoidance systems, flight management systems, and autonomous vehicles. The POMDPStressTesting.jl package is written in Julia and is part of the wider POMDPs.jl ecosystem, which provides access to simulation tools, policies, visualizations, and—most importantly—solvers. We provide different solver variants including online planning algorithms such as Monte Carlo tree search and deep reinforcement learning algorithms such as trust region policy optimization (TRPO) and proximal policy optimization (PPO). Stochastic optimization solvers such as the cross-entropy method are also available and random search is provided as a baseline. Additional solvers can easily be added by adhering to the POMDPs.jl interface.
Homepage: https://www.theoj.org/joss-papers/joss.02749/10.21105.joss.02749.pdf
Source Code: https://github.com/sisl/POMDPStressTesting.jl
Dependencies: Julia
Keywords: Adaptive stress testing; AST; Julia; Black-Box Systems; Journal of Open Source Software; reinforcement learning; POMDPs.jl ecosystem
Related Software: AdaptiveStressTesting.jl; Breach; OpenAI Gym; S-TaLiRo; Julia; POMDPs.jl
Cited in: 0 Publications

Standard Articles

1 Publication describing the Software Year