Skip to content

Latest commit

 

History

History
26 lines (16 loc) · 901 Bytes

README.md

File metadata and controls

26 lines (16 loc) · 901 Bytes

spock.jl

spock.jl is an efficient Julia implementation of the Spock algorithm for multistage risk-averse optimal control problems. It largely benefits from warm-starts and is amenable to massive parallelization.

This solver handles risk-averse optimal control problems with:

  • Linear dynamics
  • Quadratic stage and terminal costs
  • Conic risk measures
  • Convex input-state constraints

Documentation

Documentation can be found here.

Installation

Make sure to install the dependencies in the Project.toml and in your script import

include("src/spock.jl")

Examples

The examples/server_heat folder contains example code for a test system, modeling the heat of servers in a data center. The figures in the IFAC 2023 submission can be reproduced by running the scripts residuals.jl, scaling.jl and mpc_simulation.jl.