Multi-period portfolio tracking strategy based on expert uncertainty modeling and enhanced dung beetle optimization algorithm

In the face of societal uncertainties, such as financial crises and regional conflicts, which significantly impact investors’ incomes, dynamic adjustment of investment portfolios while tracking benchmarks becomes crucial for wealth maximization. This study delves into portfolio selection, aiming to maximize multi-period returns under uncertain conditions with consideration of benchmark errors. Two novel multi-period portfolio models with uncertain parameters are initially constructed based on different tracking strategies. The equivalent formulations of these models are then explored when stock returns follow linear and normal uncertain distributions. Departing from previous studies that relied on expert empirical data for asset return estimations, we introduce a data-driven parameter estimation method to determine asset distribution functions from real market data. To address the models effectively, an enhanced dung beetle optimization algorithm is proposed, integrating chaotic mapping for enhanced global search ability and an improved sine algorithm for efficient position updating. Numerical simulations across various risk levels demonstrate the effectiveness and practicality of the proposed algorithm and models. Research findings indicate investors seeking higher market returns are recommended the uncertain return tracking model; those aiming to control risks within a reliable range should refer to the uncertain risk tracking model.