Sharif Digital Repository

Evolutionary algorithms have gradually established a strong foothold as powerful search methods and have been widely applied to solve problems in many disciplines. By the way the performance of these algorithms in hierarchical applications is not satisfying. In order to improve the performance and applicability, numerous sophisticated mechanisms have been introduced and integrated into EAs. In this thesis we have tried to implement a multilayer evolutionary algorithm in order to overcome the problem. The practical results show this improvement any various aspects.

 

Self-organized critical phenomena are interesting phenomena which are ubiquitous in nature. Examples include mountain ranges , coastlines and also activities in the hu-man's brain. In these processes, without fine-tuning of any external parameter such as the temperature, the system exhibits critical behavior. In other words, the dynamics of the system, drives it towards an state in which long range correlations in space and scaling behaviors can be seen.The first successful model which could characterize such systems was BTW model, introduced by Bak , Tang and Wiesenfeld in 1987. This model, later named Abelian sandpile model, was very simple and because of this simplicity, a large amount of... 

The representation of data influences the explanation factors of data variations. Thus,the success of learner algorithms depends on the data representation. Our main contribution in this thesis is learning of high level and abstract representation using deep structure. One of the fundamental examples of representation learning is the AutoEncoders. The auto-encoder is a rigid framework that doesn’t consider explanation factors in terms of statistical concepts. So, the auto-encoders can be re-interpreted by seeing the decoder as the statistical model of interest. The role of encoder is a mechanism for inference in the model described by the decoder. Our purpose is to design such model with... 

I investigate methods of integrating prior hierarchical structure into hierarchical portfolio optimization methods. My contributions to the literature are forming a prior hierarchical structure based on investors' priorities and generating a unique representative distance matrix, which can be used as an input to other portfolio optimization methods too. In addition, I use SIC and GICs industry classifications as priory information for S&P500 companies and use them as a complementary input to the Hierarchical Risk Parity model and Hierarchical Equal Risk Contribution and compare the resultant portfolios' performance with (López de Prado, 2019)’s method of integrating prior information and... 

In this thesis, two-level optimal control of uncertain and non-uncertain large-scale systems is presented. In the proposed two-level approach, using a decomposition/coordination framework, the large-scale system is first decomposed into several interactive subsystems, at the first-level, where a smaller optimization problem is solved at each subsystem. At the second-level, a substitution type prediction method, as a coordination strategy, is used to predict the interaction between subsystems. The coordinator mainly evaluates the next update for the coordination parameters and continues to exchange information with the first-level, so that the overall optimum solution is obtained. It is shown...