Efficient Stochastic Programming in Julia

We present StochasticPrograms.jl, a user-friendly and powerful open-source framework for stochastic programming written in the Julia language. The includes both modeling tools structure-exploiting optimization algorithms. Stochastic models can be efficiently formulated using an expressive syntax, instantiated, inspected, analyzed interactively. scales seamlessly to distributed environments. Small instances of model run locally ensure correctness, whereas larger are automatically memory-efficient way onto supercomputers or clouds solved parallel These solvers based on variations classical L-shaped, progressive-hedging, quasi-gradient provide concise mathematical background various constructs available along with code listings exemplifying their usage. Both software innovations related implementation algorithmic structured highlighted. conclude by demonstrating strong scaling properties algorithms numerical benchmarks multinode setup. Summary Contribution: This paper presents implemented syntax formulating as well versatile analysis will prove useful researchers, educators, industrial users alike. Researchers benefit from readily extensible framework, which they formulate complex quickly typeset test novel Educators clean syntax. Moreover, supports theory leading textbooks. strongly believe that StochasticPrograms.jl reduce barrier entry incoming practitioners programming. Industrial make use rapidly models, analyze small locally, then large-scale production. In doing so, get capabilities free without changing access well-tested state-of-the-art implementations solvers. As is open-source, anyone these target audiences contribute new functionality framework. conclusion, providing intuitive interface extensive development environment expert users, has potential further field