Cardinality constrained portfolio optimization using multi-objective evolutionary algorithms
Constructing an optimal portfolio of assets is a multi-objective optimization process of maximizing return while minimizing risk. Two objectives need to be optimized simultaneously, making it a real world multi-objective optimization problem. Solving such problems is complex and tedious. The clas...
Saved in:
| Main Author: | |
|---|---|
| Other Authors: | |
| Format: | Final Year Project |
| Language: | English |
| Published: |
2013
|
| Subjects: | |
| Online Access: | http://hdl.handle.net/10356/54299 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Institution: | Nanyang Technological University |
| Language: | English |
| Summary: | Constructing an optimal portfolio of assets is a multi-objective optimization process of maximizing return while minimizing risk. Two objectives need to be optimized simultaneously, making it a real world multi-objective optimization problem. Solving such problems is complex and tedious.
The classical model of portfolio allocation was developed by Henry Markowitz. It talked about diversification being an important tool applied to minimize the riskiness associated with the final portfolio. However, optimization based on such mathematical derivatives is very complex and requires strong mathematical knowledge. Genetic Algorithms have a big advantage in this scenario as they offer robustness, speed and requires little mathematical knowledge.
This paper focuses on the use of two popular Multi-objective Evolutionary Algorithms to construct optimal portfolio of stocks, while being subject to a cardinality constraint. Additionally, to improve the performance of the algorithms, different techniques to filter the initial population of stocks are implemented and analyzed. |
|---|