COURSE OBJECTIVES

Whichever is your application domain, this training will provide you with an introduction to the field of nonlinear optimization and will teach you how to apply nonlinear modeling techniques to industrial applications using Artelys Knitro.

TARGET AUDIENCE

Scientists and developers interested in modeling and solving nonlinear programs using Artelys Knitro.

Nonlinear programming (NLP)

• Introduction, presentation of the training.

• Problem statement and optimality conditions.

• Newton method for unconstrained optimization. Globalization techniques.

• Interior-point and active-set methods for constrained optimization.

Solving nonlinear problems with programmatic interfaces

• Presentation, modeling and solving a nonlinear model with Artelys Knitro in Python™.

• Impact of exact versus approximate derivatives. Quasi-Newton method.

• Using Artelys Knitro in R/MATLAB®: a nonlinear least square minimization application.

Solving nonlinear problems with modeling interfaces

• Using Artelys Knitro in AMPL: modeling syntax, automatic differentiation, examples.

• Good practices in nonlinear modeling. Tips and tricks.

• Fine-tuning Artelys Knitro parameters.

• Global optimization using parallel multi-start.

Solving nonlinear models with special features

• Mixed-integer nonlinear programming (MINLP) methods. Practical example.

• Mathematical Programming with Equilibrium Constraints (MPEC). Application to computational economics and game theory.

• Convex non-smooth models.