A Neural Network Monte Carlo Approximation for Expected Utility Theory

This paper proposes an approximation method to create optimal continuous-time portfolio strategy based on a combination of neural networks and Monte Carlo, named NNMC. work is motivated by the increasing complexity models stylized facts reported in literature. We within expected utility theory for selection with constant relative risk aversion utility. The extends recursive polynomial exponential framework adopting fit value function. developed two network architectures explored several activation functions. methodology was applied four settings: 4/2 stochastic volatility (SV) model types market price risk, jumps, Ornstein–Uhlenbeck model. In only one case, closed-form solution available, which helps comparisons. report accuracy various settings terms strategy, performance computational efficiency, highlighting potential NNMC tackle complex dynamic models.