Machine-learning a family of solutions to an optimal pension investment problem

2025-11-10

AI Summary

This paper addresses the optimal pension investment decision problem under multi-parameter coupling. Methodologically, it proposes an end-to-end neural network framework that jointly models investor risk preferences, consumption preferences, and economic variables (e.g., interest rates, volatility) as inputs and directly maps them to continuous optimal investment strategies. It introduces, for the first time, a family of neural networks generalizable over the preference parameter manifold and integrates Black–Scholes modeling with duality-based verification to ensure solution reliability. The contributions are threefold: (i) enabling millisecond-scale personalized strategy generation, substantially improving computational efficiency; (ii) maintaining high accuracy—within 1% error—relative to classical numerical solutions, even in high-dimensional parameter spaces; and (iii) achieving a favorable balance among precision, interpretability, and deployability. This work establishes a novel paradigm for dynamic pension finance decision-making.

Technology Category

• Search and Optimization: Mixed Discrete/Continuous Search
• Reasoning under Uncertainty: Stochastic Optimization
• Machine Learning: Optimization

Application Category

• Economics, Online Markets and Human Computation: Incentives in network design for Web infrastructures and ecosystems
• User Modeling, Personalization and Recommendation: User privacy protection in personalized systems
• Graph Algorithms and Modeling for the Web: Graph neural networks and deep learning approaches for Web-related graphs

Abstract

We use a neural network to identify the optimal solution to a family of optimal investment problems, where the parameters determining an investor's risk and consumption preferences are given as inputs to the neural network in addition to economic variables. This is used to develop a practical tool that can be used to explore how pension outcomes vary with preference parameters. We use a Black-Scholes economic model so that we may validate the accuracy of network using a classical and provably convergent numerical method developed using the duality approach.

Problem

• Machine learning identifies optimal pension investment solutions
• Neural networks model investor risk and consumption preferences
• Validating accuracy using Black-Scholes economic model duality

Innovation

• Neural network identifies optimal pension investment solutions
• Inputs include risk preferences and economic variables
• Black-Scholes model validates network accuracy numerically