Repository navigation
ojAlgo v57.4.0, and updated integrations for COPT, CPLEX, Gurobi, MOSEK and Xpress #698
apete
started this conversation in
General, ojAlgo and related code
Replies: 0 comments
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment
Uh oh!
There was an error while loading. Please reload this page.
ojAlgo v57.4.0, and updated integrations for COPT, CPLEX, Gurobi, MOSEK and Xpress
ojAlgo v57.4.0 – pure Java
ojAlgo v57.4.0 was released on 8 October 2026. For a minor release it contains a lot: the optimisation solvers, LP, QP and MIP alike, are significantly faster and handle edge cases much better. So much has changed that all published benchmarks, even the very recent ones, will be redone.
ojAlgo has the best pure Java solvers available – we benchmark against every alternative we come across. This release widens the gap to other pure Java solvers, and narrows the gap to the best open source native solvers. Release notes: CHANGELOG.
The top commercial solvers
Still, for some models, MIP models in particular, the gap to the top commercial solvers is significant. ojAlgo's modelling layer,
ExpressionsBasedModel, does not encode which solver runs the model, so if you need a more capable solver there is an upgrade path that doesn't involve rewriting anything.Released today: updated integrations for all five, on Maven Central.
org.ojalgo:ojalgo-copt:1.1.0SolverCOPTorg.ojalgo:ojalgo-cplex:3.2.0SolverCPLEXorg.ojalgo:ojalgo-gurobi:3.2.0SolverGurobiorg.ojalgo:ojalgo-mosek:3.3.0SolverMosekorg.ojalgo:ojalgo-xpress:1.1.0SolverXpressThese versions are built against the current solver releases. New builds happen when subscribers request them – an Optimatika Subscription also gives access to the integrations' source code and support.
In code, one line:
Nothing else changes; the models you already have now run on that solver. You do also need the solver itself, installed and licensed from its vendor – the integration is not a solver licence.
They really are the best solvers available. But native binaries, installation and licensing are a real hassle, and sometimes the cost is hard to justify.
A middle ground: open source native solvers, packaged for you
There are several very good open source native solvers, such as HiGHS, SCIP and Clarabel. Optimatika's Optimisation Service packages them, ready to run, in a Docker image that you deploy in your own environment. Your existing ojAlgo model can use it as a remote solver; alternatively use the dedicated Java client or the REST API. Very little work, if the top commercial solvers cost too much in money or effort.
Getting started: Using the Optimisation Service
All available solver integrations, open source ones included: optimatika.se/solver-integrations
Questions, or a solver you'd like an integration for? Ask below.
All reactions