What is it about?
Digital assets such as cryptocurrencies are now widely used, but building reliable investment portfolios with them is still difficult. Their prices can change very quickly, they behave differently from traditional assets, and the research on how to manage digital-asset-only portfolios is scattered across many studies. This paper brings that work together in one place. We reviewed 119 publications from 2017 to 2025 and organized them into four main groups: traditional financial methods, evolutionary and swarm-based methods, machine learning and deep learning methods, and reinforcement learning methods. We explain what these approaches do, how they have been used, what their strengths and weaknesses are, and where important gaps remain. The goal is to help researchers and practitioners better understand how digital asset portfolios are currently optimized and what still needs to be improved.
Featured Image
Photo by Shutter Speed on Unsplash
Why is it important?
Digital assets are becoming an important part of modern finance, but we still do not fully understand how to manage portfolios made only of these assets. Many existing methods were originally designed for traditional markets, such as stocks and bonds, and may not work well in crypto markets where prices are highly volatile, trading never stops, and market conditions can change very quickly. This paper is important because it brings together a scattered body of research and shows what has been tried, what works, what does not work consistently, and what is still missing. By organizing the literature and identifying key gaps, the paper gives researchers a clearer roadmap for future work and helps practitioners better understand the risks and limitations of current portfolio optimization methods in the digital asset market.
Perspectives
This publication is personally important to me because it represents the foundation of my PhD research. When I first started working on digital asset portfolio optimization, I found that the literature was scattered across finance, computer science, machine learning, and blockchain research, making it difficult to understand what had already been done and where the real gaps were. Writing this survey helped me organize the field, clarify the main research directions, and identify the challenges that still need attention. I hope this article helps other researchers avoid starting from a fragmented view of the field and gives them a clearer path for developing more robust and practical portfolio optimization methods for digital assets.
Giorgos Demosthenous
University of Cyprus
Read the Original
This page is a summary of: Portfolio Optimization Methods for the Digital Asset Market: A Comprehensive Survey, ACM Computing Surveys, May 2026, ACM (Association for Computing Machinery),
DOI: 10.1145/3819577.
You can read the full text:
Contributors
The following have contributed to this page
