Recent years have seen a growing interest among investors in the new technology of blockchain and cryptocurrencies and some early investors in this new type of digital assets have made significant gains. The heuristic algorithm, differential evolution, has been advocated as a powerful tool in portfolio optimization. We propose in this study two new approaches derived from the traditional differential evolution (DE) method: the GARCH-differential evolution (GARCH-DE) and the GARCH-differential evolution t-copula (GARCH-DE-t-copula). We then contrast these two models with DE (benchmark) in single and multi-period optimizations on a portfolio consisting of five cryptoassets under the coherent risk measure CVaR constraint. Our analysis shows that the GARCH-DE-t-copula outperforms the DE and GARCH-DE approaches in both single- and multi-period frameworks. For these notoriously volatile assets, the GARCH-DE-t-copula has shown risk-control ability, hereby confirming the ability of t-copula to capture the dependence structure in the fat tail.

中文翻译：

---

用于多周期加密货币投资组合优化的基于差分进化 copula 的方法

近年来，投资者对区块链和加密货币的新技术越来越感兴趣，这种新型数字资产的一些早期投资者取得了显着收益。启发式算法，差分进化，一直被认为是投资组合优化的有力工具。我们在这项研究中提出了两种源自传统差分进化 (DE) 方法的新方法：GARCH-差分进化 (GARCH-DE) 和 GARCH-差分进化 t-copula (GARCH-DE-t-copula)。然后，我们将这两个模型与 DE（基准）在相干风险度量 CVaR 约束下对由五个加密资产组成的投资组合的单期和多期优化进行对比。我们的分析表明，GARCH-DE-t-copula 在单周期和多周期框架中都优于 DE 和 GARCH-DE 方法。对于这些臭名昭著的波动性资产，GARCH-DE-t-copula 已经表现出风险控制能力，从而证实了 t-copula 捕捉肥尾依赖结构的能力。