Computational Intelligence and Power Systems
Keywords |
Classification |
Keyword |
OFICIAL |
Electrical and Computer Engineering |
Instance: 2024/2025 - 1S
Cycles of Study/Courses
Acronym |
No. of Students |
Study Plan |
Curricular Years |
Credits UCN |
Credits ECTS |
Contact hours |
Total Time |
PDEEC |
6 |
Syllabus |
1 |
- |
6 |
42 |
162 |
Teaching Staff - Responsibilities
Teaching language
English
Objectives
This course aims at making students familiar with a number of tools pertaining to the domain of computatinal intelligence, which will be useful in dealing with power and energy system models in their research activity, in outher courses and in their profesional life.
The students having taken this course shall be able to develop models under the compytational intelligence paradigm, to programm algorithms and to discuss their results in terms of accuracy, effort and credibility.
Learning outcomes and competences
As a result of the course, the student will be better prepared to use Computational Intelligence methods.
Working method
Presencial
Program
The course syllabus includes:
- Evolutionary Computing: history, basic concepts; self adaptive models; convergence; theoretical foundations. Variants: genetic algorithms, evolutionary programming/evolution strategies.
- Particle Swarm Optimization: mouvement equation, convergence; convergence control, constriction.
- Evolutionary Particle Swarms: self-adaptive recombination.
- Cross-entropy optimization.
- Application of evolutionary swarm and cross-entropy algorithms to power system problems and discussion of the results. Application in reliability: population-based methods.
- Concept of mappers. Criticism of Minimum Square Error (MSE) as a training cost function. Introduction to Information Theoretic Learning (ITL). Training mappers under an Entropy-related MEE cost criterium. Correntropy and MCC criterium. Application of mappers in power system problems.
- Autoencoders and deep networks. Data compression and feature reduction. Interpretation in the information flow context. Unsupervised training using Infornation Theoretic concepts.
- Clustering using ITL concepts. Renyi Entropy and Cauchy-Schwarz divergence. Kulback-Leibler divergence. Mean Shift algorithms and Information Theoretic Mean Shift.
- Polynomial networks: introduction to GDHM (Group Data Handling Method). Construction/training algorithm. Application of GDHM in power system problems.
Mandatory literature
Vladimiro Miranda; REDESIGNING MODELS WITH INFORMATION THEORETIC LEARNING CONCEPTS: A SHORT OVERVIEW, 2018 (Text supplied by the lecturer)
Teaching methods and learning activities
Lectures and assignments.
keywords
Technological sciences > Engineering > Electrical engineering
Technological sciences > Engineering > Systems engineering
Physical sciences > Mathematics > Algorithms
Evaluation Type
Distributed evaluation with final exam
Assessment Components
Designation |
Weight (%) |
Apresentação/discussão de um trabalho científico |
20,00 |
Exame |
50,00 |
Trabalho prático ou de projeto |
30,00 |
Total: |
100,00 |
Amount of time allocated to each course unit
Designation |
Time (hours) |
Apresentação/discussão de um trabalho científico |
10,00 |
Elaboração de projeto |
40,00 |
Frequência das aulas |
50,00 |
Total: |
100,00 |
Eligibility for exams
In order to have his/her attendance of the course recognized, the student must complete all assignments.
Calculation formula of final grade
In order to obtain a passing mark, the student must complete all the assignments with a positive evaluation and have a minimum of 8/20 in the final exam.
To compute the final mark, the exam will entre with a weight of 50% and the set of assignments with a weight of 50%.
Special assessment (TE, DA, ...)
No exceptions allowed to the general evaluation method.
Classification improvement
In order to improve the evaluation results, the student may be submitted to a second exam and may request a period to improve one of the assignments.