|
Abstract
|
|
|---|---|
| Examples of broadly known evolutionary computation paradigms are Genetic Algorithms (GAs) and Estimation Distribution Algorithms (EDAs). The main difference between these models is the evolution process: In GAs this is based on crossover and mutation operators, without explicit expression of the characteristics of selected individuals within the population. EDAs take these explicit characteristics into account by considering the interdependencies between the variables that form an individual and by learning a probabilistic graphical model that represents them. This thesis analyses the importance of two aspects in the overall performance: the fitness of each individual when it is evaluated using the fitness function and analyse the most appropriate way to represent a problem for its solving by EDAs. | |
|
International
|
No |
|
|
|
|
Type
|
Doctoral |
|
Mark Rating
|
Sobresaliente cum laude |
|
Date
|
20/05/2010 |