Hybrid Metaheuristic Design for Solving the Cardinality-Constrained Portfolio Problem: Comparative Analysis in Six Equity Markets

We present a hybrid metaheuristic for mean-variance portfolio optimization with a cardinality constraint. The algorithm combines Continuous Ant Colony Optimization (CACO), a Genetic Algorithm (GA), and Artificial Bee Colony (ABC), with elitism and localized mutation to balance exploration and exploitation. Across six benchmark equity markets (S&P ۱۰۰, FTSE ۱۰۰, Hang Seng, Nikkei ۲۲۵, DAX ۱۰۰, XU۱۰۰/XU۰۳۰), the method yields lower mean/variance return error and smaller mean Euclidean distance to the unconstrained efficient frontier than competing settings. On S&P ۱۰۰, the best configuration achieves VRE ۲.۵۲, MRE≈۰.۷۰, and MEUCD ≈۱e-۴. The approach remains competitive on more volatile markets (e.g., DAX, Nikkei), indicating robustness.

کلیدواژه ها:

Portfolio optimization ، cardinality constraint ، metaheuristics ، ant colony optimization ، genetic algorithms ، artificial bee colony