Machine-learning a family of solutions to an optimal pension investment problem
Updated
Updated · arxiv.org · Aug 20
Machine-learning a family of solutions to an optimal pension investment problem
1 articles · Updated · arxiv.org · Aug 20
Summary
Researchers have developed a neural network method to solve a family of optimal pension investment problems, varying key investor preference parameters.
Their main advance is a loss function scaling technique that enables effective training across diverse risk and consumption preferences, with proven algorithmic convergence.
Validated against classical numerical methods, this tool allows pension professionals to explore how outcomes change with different preference settings, aiding tailored pension design.