rcss: Subgradient and duality approach for dynamic programming

TL;DR

This paper introduces the rcss R package, which approximates value functions in dynamic programming using convex piecewise linear functions and evaluates their quality with a pathwise method.

Contribution

It presents a novel implementation of convex piecewise linear approximation for value functions in dynamic programming within an R package.

Findings

Provides an accessible tool for value function approximation

Demonstrates the effectiveness of convex piecewise linear methods

Includes a method for assessing approximation quality

Abstract

This short paper gives an introduction to the \emph{rcss} package. The R package \emph{rcss} provides users with a tool to approximate the value functions in the Bellman recursion using convex piecewise linear functions formed using operations on tangents. A pathwise method is then used to gauge the quality of the numerical results.