A Two-Stage Decision Support System for Sustainability-Aware Long–Short Portfolio Optimization

Battle Count: 75

Highly relevant to quantitative trading and portfolio management. The paper addresses long-short portfolio construction with ESG integration, uses a non-convex optimization framework (Omega ratio), and proposes an advanced metaheuristic solver (AMPSO). The two-stage architecture (screening + optimization) mirrors institutional investment processes. The regime-aware classification and adaptive optimization are directly applicable to systematic trading strategies. However, the focus on monthly rebalancing and the absence of transaction cost modeling somewhat limit direct applicability to high-frequency or intraday trading.

High implementation complexity due to: (1) multi-criteria decision making pipeline with MEREC weighting, TODIM dominance calculations, and TODIMSort classification with dynamic limiting profiles; (2) market regime detection using multiple technical indicators (SMA, ADX, RSI, MACD); (3) AMPSO solver with adaptive operator selection, multiple crossover/mutation operators, and projection-based constraint repair; (4) ESG data integration and regime-conditioned empirical distributions; (5) careful handling of regulatory constraints (short-selling bans). The mathematical formulations are well-defined but require significant coding effort, particularly the O(n) projection algorithm and the reinforcement-learning-based operator selection.

The paper provides detailed parameter settings (Table 5), full pseudocode for AMPSO (Algorithm 1), operator definitions in Supplementary Material, and clear mathematical formulations. However, no GitHub repository or code link is provided. Data comes from Refinitiv (proprietary). The methodology is well-documented but implementation requires significant effort due to the complexity of the two-stage pipeline and multiple operators.