Information Technology Reference
In-Depth Information
Fig. 4. Annualized return for textile in testing
5
Conclusions
In our proposed MOGA here, we mainly use NSGA-II to study the effect of the two
competitive objectives of return and risk, and then refine the method with the assis-
tance of relevant domain knowledge from investment. Especially, we reset the scores
of the portfolios with low risk but negative return to zero to prevent them from being
selected and assign higher weights to the portfolios of higher return per unit of risk to
construct more profitable portfolios. The experimental results showed that the MOGA
method significantly outperformed the benchmark. In the future, we intend to employ
other sophisticated versions of the MOGA, such as the MOEA/D to substitute the
simple weight-sum MOGA used in this study. Furthermore, we will attempt to gene-
ralize the proposed model here to other investment problems such as asset allocation.
References
1. Bermudez, J.D., Segura, J.V., Vercher, E.: A Fuzzy Ranking Strategy for Portfolio Selec-
tion Applied to the Spanish Stock Market. In: The 2007 IEEE International Conference on
Fuzzy Systems, pp. 1-4 (2007)
2. Chen, S.S., Huang, C.F., Hong, T.P.: A Multi-objective Genetic Model for Stock Selec-
tion. In: The 27th Annual Conference of the Japanese Society for Artificial Intelligence
(2013)
3. Chiang, C.H.: Using Genetic Algorithms to Find Fundamental Selection and Technical
Timing Rules. Master Thesis, National Central University (2002)
4. Deb, K., Pratap, A., Agarwal, S., Meyarivan, T.: A Fast and Elitist Multiobjective Genetic
Algorithm: NSGA-II. The 2002 IEEE Transactions on Evolutionary Computation 6(2),
182-197 (2002)
5. Hoklie, Z.L.R.: Resolving Multi Objective Stock Portfolio Optimization Problem Using
Genetic Algorithm. In: The Second International Conference on Computer and Automa-
tion Engineering, vol. 2, pp. 40-44 (2010)
Search WWH ::




Custom Search