Near Optimal Hamiltonian-Control and Learning via Chattering

TL;DR

This paper introduces a chattering algorithm that efficiently learns near-optimal control policies for complex non-linear problems, demonstrated through real-time enterprise scheduling applications.

Contribution

It presents a novel, simple chattering approach that reduces complex control problems to linear programs for near-optimal solutions.

Findings

Algorithm effectively learns near-optimal policies

Reduces control problems to linear optimization

Successfully applied in real-time enterprise scheduling

Abstract

Many applications require solving non-linear control problems that are classically not well behaved. This paper develops a simple and efficient chattering algorithm that learns near optimal decision policies through an open-loop feedback strategy. The optimal control problem reduces to a series of linear optimization programs that can be easily solved to recover a relaxed optimal trajectory. This algorithm is implemented on a real-time enterprise scheduling and control process.