2024

Numerical Optimization

Name: Numerical Optimization
Code: MAT13652M
6 ECTS
Duration: 15 weeks/156 hours
Scientific Area: Mathematics

Teaching languages: Portuguese
Languages of tutoring support: Portuguese

Sustainable Development Goals

Learning Goals

To give a solid training on numerical methods of Optimization in terms of theory and applications. The student will be able to choose the best algorithm to solve practical real life problems with specific software. It is intended too that the student acquires the competence to have a critical view on the models and software limitations.

Contents

Classical theory of optimization. Necessary and sufficient extreme conditions. Convex sets and functions.
Numerical optimization methods for unrestricted one and multiple variable functions.
Constrained numerical optimization methods. Penalty functions. Interior point method.
Introduction to genetic and evolutionary algorithms.
Implementation of algorithms in Python. Software for numerical optimization (SciPy package).

Teaching Methods

Structured exposition with application emphasis, exercise solving, implementation of some algorithms em Python.

Assessment

The students can opt for continuous avaliation or the final exam.
Continuous avaluation is privileged and consists of a final test on the covered contents (50%) and a individual (or in small group) computational project presented publicly (50%).
At the evaluation by exam, it is necessary to use computational tools to solve some questions.
To obtain approval for the CU, the student must have at least 10 values in the final grade.