NumFOCUS small development grant: adding complex number support to JuMPannouncements ·
The JuMP Steering Committee is pleased to announce that we have received a small development grant from NumFOCUS to add complex number support to JuMP.
Currently, JuMP and MathOptInterface are limited to formulating optimization problems with real-valued decision variables. However, complex numbers appear in a variety of industrial optimization problems. One of the most important of these is the AC optimal power flow problem (AC-OPF), which is used world-wide to optimize the operations of electrical power grids. Currently, users in JuMP are forced to decompose their problem into the real and complex parts. This leads to a duplication in which twice as many variables are needed, reducing readability and increasing the likelihood of bugs. As a result, complex number support has been a common feature request for users of JuMP.
The person being funded is Dr. Benoît Legat, a core contributor to JuMP, and the project builds upon his work developing ComplexOptInterface.jl.
The grant has three key deliverables:
- Improve ComplexOptInterface.jl from a proof-of-concept into a robust, production-ready library
- Integrate ComplexOptInterface into MathOptInterface, test and document
- Integrate ComplexOptInterface into JuMP, test and document.
Benoît will be working on a part-time basis over the coming year, and will present an update during JuMP-dev 2022.
Miles Lubin, Juan Pablo Vielma, Carleton Coffrin, Oscar Dowson and Changhyun Kwon.