군집 지능 알고리즘을 활용한 포트폴리오 연구

Abstract

While metaheuristics have profoundly impacted various fields, domestic financial portfolio optimization research, particularly in asset allocation, remains underdeveloped. This study investigates metaheuristic algorithms for investment strategy optimization. Results reveal that metaheuristic-optimized portfolios outperform the Dow Jones Index in Sharpe ratios, highlighting their potential to significantly enhance risk-adjusted returns. A comparative analysis of Ant Colony Optimization (ACO) and Cuckoo Search Algorithm (CSA) shows CSA's slight superiority in risk-adjusted performance. This advantage is attributed to CSA's maintained randomness and Lévy flight model, which effectively balance local and global search, whereas ACO may converge prematurely due to path reinforcement. These findings underscore metaheuristics' capacity to maximize expected returns at given risk levels, offering flexible, robust solutions for investment strategy optimization.