Unlocking the Power of Google OR-Tools with MathOptInterface.jl
Abstract
JuMP and MathOptInterface.jl give access to many solvers, both very common in the industry and more specialised. Google offers its own in-house solvers as part of the open-source package OR-Tools: Glop, a simplex solver; CP-SAT, an award-winning constraint-programming solver; PDLP, a first-order solver for large-scale linear programming.
ORTools.jl is a recent package that gives access to these solvers through MathOptInterface.jl. It supports both local and remote use, meaning that users do not need a local installation to solve linear and integer problems thanks to Google Cloud. More recently, ORTools.jl started offering a native interface for constraint programming building upon the work in MathOptInterface.jl.
However, OR-Tools have more than this to offer, including a scalable routing solver for large-scale VRPs or state-of-the-art set-cover solver. MathOptInterface.jl does not yet propose an interface for these problems and we would like to gauge the community’s interest in these specific solvers.
ORTools.jl is a recent package that gives access to these solvers through MathOptInterface.jl. It supports both local and remote use, meaning that users do not need a local installation to solve linear and integer problems thanks to Google Cloud. More recently, ORTools.jl started offering a native interface for constraint programming building upon the work in MathOptInterface.jl.
However, OR-Tools have more than this to offer, including a scalable routing solver for large-scale VRPs or state-of-the-art set-cover solver. MathOptInterface.jl does not yet propose an interface for these problems and we would like to gauge the community’s interest in these specific solvers.
