<![CDATA[Traditional mean-variance portfolio optimization often faces limitation in accommodating practical investment constraints, particularly in restricting the number of selected assets while ensuring a minimum expected return. To address these limitations, this study proposes a Genetic Algorithm (GA)-based portfolio optimization framework that simultaneously incorporates cardinality and minimum return constraints for constructing efficient portfolios. The main contribution of this study is the integration of these practical investment constraints into a GA framework to improve portfolio optimization under realistic investment scenarios using LQ45 stocks. Historical daily stock price data of LQ45 constituents were collected from Yahoo Finance and used to estimate portfolio returns, portfolio risk (standard deviation), and Sharpe ratios. The optimization process minimizes portfolio risk while enforcing cardinality and minimum return constraints through penalty function. Experimental results show that the proposed framework successfully generated efficient portfolios under various parameter settings. The best portfolio was obtained using a cardinality constraint of three stocks and a minimum return target of 10%. This portfolio achieved a return of 14.17%, a risk (standard deviation) of 0.1670, and a Sharpe ratio of 0.4867. In comparation, the equal-weighted LQ45 portfolio produces a return of -16.64% with a Sharpe ratio of -1.2710, indicating substantially lower investment performance. These findings demonstrate that the proposed GA-based framework provides an effective and practical portfolio optimization strategy under realistic investment constraints while highlighting the potential of evolutionary optimization methods for improving risk-adjusted portfolio performance in dynamic market environments.]]>
Copyrights © 2026
