StochJuMP.jl

The StochJuMP.jl package provides a scalable algebraic modeling framework for stochastic programming problems in Julia. StochJuMP.jl is an extension of the JuMP.jl package, which is as fast as AMPL and faster than any other modeling tools such as GAMS and Pyomo (see this).

An example of the StochJuMP.jl package reads:

using StochJuMP

numScen = 2
m = StochasticModel(numScen)

@defVar(m, 0 <= x <= 1)
@defVar(m, 0 <= y <= 1)

@addConstraint(m, x + y == 1)
setObjective(m, :Min, x*x + y)

for i in 1:numScen
    bl = StochasticBlock(m)
    @defVar(bl, w >= 0)
    @addConstraint(bl, w - x - y <= 1)
    setObjective(bl, :Min, w*w + w)
end

Solvers for StochJuMP

The StochJuMP model can be solved by either PIPS or DSP. PIPS is an open-source parallel interior point solver for stochastic convex and nonconvex continuous programs. DSP is a open-source package of the parallel decomposition methdos for stochastic mixed-integer programs. The Julia interface for PIPS and DSP are also available in PIPS.jl and DSPsolver.jl, respectively.