English

Multi-objective dynamic programming with limited precision

Machine Learning 2020-09-18 v1 Machine Learning

Abstract

This paper addresses the problem of approximating the set of all solutions for Multi-objective Markov Decision Processes. We show that in the vast majority of interesting cases, the number of solutions is exponential or even infinite. In order to overcome this difficulty we propose to approximate the set of all solutions by means of a limited precision approach based on White's multi-objective value-iteration dynamic programming algorithm. We prove that the number of calculated solutions is tractable and show experimentally that the solutions obtained are a good approximation of the true Pareto front.

Keywords

Cite

@article{arxiv.2009.08198,
  title  = {Multi-objective dynamic programming with limited precision},
  author = {L. Mandow and J. L. Pérez de la Cruz and N. Pozas},
  journal= {arXiv preprint arXiv:2009.08198},
  year   = {2020}
}