New release: LocalSolver 11.5

We are pleased to announce the release of LocalSolver 11.5. Discover below the main novelties coming with this new release.

Vehicle Routing problems

Vehicle Routing problems are solved even faster. The “gap” mentioned below is the relative gap in % between the solutions computed by LocalSolver 11.5 within 1 minute of running time on a standard server (Intel Core i7-7700K processor, 4 cores, 4.2 GHz, 8MB cache, 64GB RAM) and the best-known solutions available in the research literature (computed using dedicated algorithms, within arbitrary long running times, and possibly taking advantage of more powerful hardware).

LocalSolver 11.5 improves its lower bounds for the Vehicle Routing Problem with Time Windows (CVRPTW) with an average dual gap of 9% for instances with up to 100 deliveries within 1 minute of running time.

Production Scheduling problems

The “gap” mentioned below is computed in the same way as above. In particular, LocalSolver 11.5 is launched with a time limit of 1 minute only.

Global optimization capabilities

  • Mixed-Integer Nonlinear optimization (MINLP): optimal solutions found in less than 1 minute of running time for 734 instances over the 1,668 in the MINLPLib benchmark (+38 instances compared to LocalSolver 11.0, +94 instances compared to LocalSolver 10.5). Note that LocalSolver checks constraint feasibility with a relative 1e-6 numerical precision and objective optimality with a relative 1e-4 numerical precision.

Simulation optimization capabilities

  • Constrained black-box optimization: 100% of feasible solutions and 77% of optimal solutions found with an average of 50 black-box evaluations in 5 runs over 36 instances from the research literature, compared to 87% (resp. 82%) of feasible solutions and 31% (resp. 21%) of optimal solutions found with an average of 112 (resp. 124) black-box evaluations for RBFOpt (resp. NOMAD) black-box optimization solver.

Modeling APIs and language (LSP)

use csv;

function input() {
    data = csv.parse("./data.csv");
    for [row in data.rows] {
        //...   
    }
}
  • LocalSolver modeling language (LSP) can be used as a plain programming language with the addition of a main function.
function main() {
    readData();
    solveSubProblem1();
    solveSubProblem2();
    writeSolution(solution);
}

LocalSolver Studio

This new version comes with a new product: LocalSolver Studio. It will be launched separately in a Beta version in the following weeks. LocalSolver Studio allows you to develop and test LocalSolver-powered optimization applications anywhere from your browser. On a single interface, you benefit from:

  • An online disk space to import and share your data;
  • An editor to write, launch, and debug your LocalSolver optimization models by taking advantage of the power of LocalSolver Cloud;
  • An interactive dashboard to bring your LocalSolver optimization models to life in the form of KPIs, maps, Gantt charts, and other various GUI widgets.

Stay tuned!

Try LocalSolver Studio exclusively.

LocalSolver Studio

New code examples

New code examples are available in the Example Tour to help you get started quickly with LocalSolver:

We are at your disposal to accompany you in the discovery of LocalSolver 11.5. Please ask for your free trial license by registering here. In the meantime, don’t hesitate to contact us for further information or support.

Share