StochasticBarrier.jl
Toolbox for stochastic barrier functions in noisy systems
StochasticBarrier.jl, developed in collaboration with Rayan Mazouz, is a Julia toolbox for generating stochastic barrier functions for discrete-time stochastic systems with additive Gaussian noise.
Features
- Supports linear, polynomial, and piecewise-affine uncertain dynamics
- Includes Sum-of-Squares (SOS) and piecewise-constant (PWC) approaches
- PWC engines based on dual LP, CEGS LP, and projected gradient descent
- Benchmark scripts used to reproduce experiments from the toolbox paper
Applications
- Safety verification of stochastic systems via barrier certificates
- Reproducible benchmarking workflows for SOS and PWC methods
- Experiment pipelines with Docker-based setup