| Abstrak/Abstract |
The non-convex Risk Parity (RP) portfolio opti
mization presents challenges due to the potential presence of
multiple local minima, making it difficult to identify the optimal
solution. Meta-heuristic algorithms, known for their flexibility,
are ideal for addressing this issue, as they effectively balance
exploration of new solution spaces with refinement of promising
candidates. This study compares the performance of three
meta-heuristic algorithms— Genetic Algorithm (GA), Particle
Swarm Optimization (PSO), and Ant Colony Optimization
for continuous domains (ACOR)—in solving the non-convex
RP portfolio optimization problem. Using both real-world and
simulated datasets, the first empirical study demonstrates the
superior performance of PSO. A second study, employing the
rolling-window method, evaluates the RP portfolio against the
Equally Weighted (EW) and Global Minimum Variance (GMV)
portfolios. The results show that, while the RP portfolio does
not consistently outperform the others across all metrics, it
excels in minimizing Maximum Drawdown (MD) and Value
at-Risk (VaR). This research contributes to the literature by
offering a thorough comparison of meta-heuristic algorithms
for non-convex RP portfolio optimization and highlighting the
RP portfolio’s robustness in risk management. |
