<ul><li>Portfolio optimization involves selecting asset weights to minimize a risk-reward objective, such as portfolio variance.</li><li>Sparse portfolio selection imposes a restriction where only a limited number of assets can be included from a larger pool.</li><li>Proposed scalable gradient-based approach transforms the sparse selection problem into a constrained continuous optimization task using Boolean relaxation.</li><li>The algorithm allows for stable convex starting points, controlled progression towards a sparse binary solution, and matches commercial solvers' results in most cases.</li></ul>

A Scalable Gradient-Based Optimization Framework for Sparse Minimum-Variance Portfolio Selection

Discover more