This version comes with a redesign of the C API improving flexibility when manipulating your models. The user can directly provide the constraint structure (e.g. linear, quadratic, quadratic conic) to the solver without using callback functions. It is also possible to provide derivatives for only a subset of the model, Knitro being able to automatically approximate the missing derivatives.
Artelys Knitro intègre désormais une spécialisation algorithmique permettant de gérer efficacement les contraintes quadratiques coniques, présentes dans les modèles SOCP que l’on retrouve dans de nombreuses applications (l’économie, la physique, la finance …). Cette évolution majeure pour les problèmes SOCP s’accompagne d’une amélioration notable des performances sur les problèmes QCQP.
• CG preconditioning now available for all classes of problems
• New parallel linear solvers (HSL MA86 and MA97) allowing speedups on large scale problems
• The Knitro R interface is now open source
• Several enhancements of the MINLP and active-set algorithms
• General performance improvements in particular for convex problems
— We are pleased to announce that Artelys Knitro 13.1 is now available! Check out its improvements, especially on mixed-integer nonlinear problems (MINLP).
Artelys becomes a partner of the Grand Est Region thanks to its expertise in artificial intelligence!
— Artelys has been referenced by the Grand Est Region to assess the artificial intelligence maturity of companies in the region.
— Artelys was thrilled to organize the METIS 2 Dissemination webinar on May 31. In the course of this event, we had the opportunity to present the latest developments of the EU energy system model METIS and to discuss the future of EU energy system models with representatives from TSOs and industry.
— Artelys, in cooperation with Trinomics, prepared a study evaluating flexibility options to support decarbonisation for the Energy Community.
subscribe to our newsletters