In a controlled lifecycle benchmark, per-date backward neural policies beat single and two-regime networks on welfare and Bellman residuals, and a shape penalty fixes negative MPC violations.
1957.Dynamic Programming
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
math.OC 1years
2026 1verdicts
CONDITIONAL 1roles
background 1polarities
background 1representative citing papers
citing papers explorer
-
Simulation-Based Neural Policies for Portfolio Choice: Architecture, Training, and Interpretability
In a controlled lifecycle benchmark, per-date backward neural policies beat single and two-regime networks on welfare and Bellman residuals, and a shape penalty fixes negative MPC violations.