Sharpe Based Portfolio Optimization Using Slime Mould Algorithm

The bio-inspired metaheuristics have evolved as effective alternatives to handle portfolio optimization problems, which remain a key instrument in investment analysis. This study proposes the adoption of the Slime Mould Algorithm (SMA) to address the portfolio optimization problem with the objective of maximizing the Sharpe ratio under Lintner’s framework of short selling. The SMA-based portfolio optimization is performed on two distinct datasets from the NASDAQ and the NSE, with optimal parameter settings fine-tuned using Sobol’s sensitivity analysis and random and grid search methods. Rigorous benchmarking experimentation against three widely used metaheuristics, the Genetic Algorithm (GA), the Particle Swarm Optimization (PSO), and the Differential Evolution (DE), over multiple independent runs, demonstrates the performance superiority of SMA over others. Statistical tests like the KruskalWallis H test and the post-hoc Dunn’s test also confirm a significant performance difference among the four algorithms. The study concludes the reliability and effectiveness of SMA in Sharpe-based portfolio optimization.