AI Paper Atlas

Research map / Reinforcement learning / MDPs and planning

Multi-Objective Markov Decision Processes for Data-Driven Decision Support

Daniel J. Lizotte, Eric B. Laber

JMLR 2016 · Reinforcement learning › MDPs and planning · approach: Robust MDP policy gradient · 51 citations (OpenAlex)

Read the paperCheck your idea against 137,217 papers

AI-extracted. The measured effects below were read by AI from the paper's full text (no summary yet). Numbers and claims can be wrong or missing context, so check the paper before you cite or build on it.

Measured effects reported in the paper

What was changedWhereMeasuredEffect
use non-deterministic fitted-Q with Pareto domination instead of convex-combination scalarizationfitted-Q for CATIE · CATIE Phase 1mean choices per state: 2.26 -> 4.14Improved
use Strong Practical Domination instead of Pareto dominationnon-deterministic fitted-Q · CATIE Phase 1mean choices per state: 4.14 -> 4.30Improved
use Strong Practical Domination instead of Pareto dominationnon-deterministic fitted-Q · CATIE Phase 1percent of states with an action eliminated: 68% -> 55%Improved
use strict future-consistency definition instead of the weaker onenon-deterministic fitted-Q with Strong Practical Domination · CATIE Phase 1mean choices per state: 4.30 -> 2.56Hurt
use strict future-consistency definition instead of the weaker onenon-deterministic fitted-Q with Strong Practical Domination · CATIE Phase 1percent of states with an action eliminated: 55% -> 100%Hurt
use phi-consistency to prune policiesnon-deterministic fitted-Q, Phase 2 Efficacy · CATIE Phase 1number of candidate policies: 1 048 576 -> 98Improved

Closest papers by meaning

Writing a paper in this area? scime draft is a free LaTeX editor with AI included, real-time co-authors and a check of your references and venue rules before you submit. It also imports OpenAI Prism projects.

Open scime draft

Get the next scime Dispatch and a monthly field update

Free. We send a confirmation link first; every email has a one-click unsubscribe. Privacy