Generalizes the Tsitsiklis-van Roy error bound for aggregation in discounted DP to soft and feature-based schemes.
Feature-based methods for large scale dynamic programming
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
math.OC 1years
2025 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
An Error Bound for Aggregation in Approximate Dynamic Programming
Generalizes the Tsitsiklis-van Roy error bound for aggregation in discounted DP to soft and feature-based schemes.