Optimization and solvers
What is Artelys Knitro? An introduction to Artelys's nonlinear optimization solver
Artelys Knitro is a solver for nonlinear optimization. It offers NLP and MINLP algorithms for business and industry applications. It is used on problems such as portfolio optimization, optimal power flow, nonlinear model predictive control and Nash equilibrium problems.
- Tool
- Artelys Knitro
- Made by
- Artelys
- Official site
- artelys.com/solvers/knitro
- Category
- Optimization and solvers
Who makes Artelys Knitro
Knitro is developed by Artelys, an independent company specialised in optimization, decision support and modeling. Artelys was founded in 2000 in Paris.
The official Knitro website is linked below this page.
Official website: artelys.com/solvers/knitro
What Artelys Knitro is for
Knitro is for solving nonlinear optimization problems. In these problems, the objective or the constraints are nonlinear functions of the decision variables. Some also include integer decisions, which is the setting for MINLP (mixed-integer nonlinear programming) algorithms.
The vendor points to problem types such as portfolio optimization, optimal power flow, nonlinear model predictive control and Nash equilibrium problems. These come from finance, energy, control engineering and game-theoretic modeling.
Who Artelys Knitro is for
The vendor names three audiences: optimization practitioners, academic institutions, and business and industry users. In practice, that means people who already formulate problems as mathematical programs and need a solver to compute solutions, rather than people looking for a ready-made dashboard.
Knitro is supported on Windows, Linux and Mac, so it can be used on the operating systems that practitioners and institutions commonly work with.
Notable capabilities
Knitro combines a range of algorithms with features that help set up, run and tune a solve. The points below summarize what the vendor documents.
- Algorithm range: Knitro offers four NLP algorithms and three MINLP algorithms. The NLP set includes two interior-point/barrier algorithms and two active-set/SQP algorithms.
- MINLP techniques: for mixed-integer problems it includes heuristics, cutting planes and branching rules.
- Complementarity constraints: a special routine handles problems that contain complementarity constraints.
- Global search and concurrency: a parallel multistart feature supports global optimization, and multiple algorithms can run concurrently.
- Version 16.0: this release introduced a native Nonlinear Modeler and a Concurrent Solver.
- Automatic tuning: Knitro offers automatic and parallel tuning of option settings.
- Derivatives: it automatically computes approximate first- and second-order derivatives.
- Initialization and infeasibility: it provides smart initialization and fast infeasibility detection.
How Artelys Knitro fits optimization and solvers
Within the category of optimization and solvers, Knitro sits at the computational core. A solver does not decide what the objective should be or which constraints matter. It takes a model that someone has defined and searches for a solution that satisfies it. Knitro's focus is the nonlinear case, with continuous and mixed-integer variants.
For decision intelligence, decision support and decision analytics, this kind of tool is typically one component of a larger workflow. Analysts or engineers build a model of a decision, such as how to allocate a portfolio or how to dispatch power, and a solver produces the recommended values. Those results can then feed reports, scenario reviews or other decision-support systems. Artelys describes itself as specialised in optimization, decision support and modeling, which is consistent with that role.
Features such as infeasibility detection and automatic option tuning are relevant here because real decision models are often imperfect at first. Learning quickly that a formulation has no feasible solution, or reducing the manual effort of configuring a solve, helps people refine a model before relying on its output.
Frequently asked questions
What is Artelys Knitro used for?
Knitro is used to solve nonlinear optimization problems. Examples named by the vendor include portfolio optimization, optimal power flow, nonlinear model predictive control and Nash equilibrium problems.
Who develops Artelys Knitro?
Knitro is developed by Artelys, an independent company specialised in optimization, decision support and modeling. Artelys was founded in 2000 in Paris.
Does Artelys Knitro support MINLP problems?
Yes. Knitro offers three MINLP algorithms, along with heuristics, cutting planes and branching rules for mixed-integer nonlinear problems.
Which operating systems does Artelys Knitro support?
Knitro is supported on Windows, Linux and Mac.
What is new in Knitro 16.0?
Knitro 16.0 introduced a native Nonlinear Modeler and a Concurrent Solver.
Reasonelm is independent and is not affiliated with Artelys. Product names belong to their owners. Features change; confirm details on the official website.