A practical guide to multi-objective reinforcement learning and planning

Autor(es):Mostrar menos -

Hayes, Conor F. ; Radulescu, Roxana ; Bargiacchi, Eugenio ; Kallstrom, Johan ; Macfarlane, Matthew ; Reymond, Mathieu ; Verstraeten, Timothy ; Zintgraf, Luisa M. ; Dazeley, Richard ; Heintz, Fredrik ; Howley, Enda ; Irissappane, Athirai A. ; Mannion, Patrick ; Nowe, Ann ; Ramos, Gabriel ; Restelli, Marcello ; Vamplew, Peter ; Roijers, Diederik M.

• Tipo de documento: Artigo Científico
• Fonte: AUTONOMOUS AGENTS AND MULTI-AGENT SYSTEMS; v. 36, n. 1, p. 59-pg., 2022-04-01.

Resumo

Real-world sequential decision-making tasks are generally complex, requiring trade-offs between multiple, often conflicting, objectives. Despite this, the majority of research in reinforcement learning and decision-theoretic planning either assumes only a single objective, or that multiple objectives can be adequately handled via a simple linear combination. Such approaches may oversimplify the underlying problem and hence produce suboptimal results. This paper serves as a guide to the application of multi-objective methods to difficult problems, and is aimed at researchers who are already familiar with single-objective reinforcement learning and planning methods who wish to adopt a multi-objective perspective on their research, as well as practitioners who encounter multi-objective decision problems in practice. It identifies the factors that may influence the nature of the desired solution, and illustrates by example how these influence the design of multi-objective decision-making systems for complex problems. (AU)

• Processo FAPESP: 20/05165-1 - Comunicação e aprendizado de máquina em mobilidade urbana: uma abordagem multiagente e multiobjetivo
• Beneficiário: Ana Lúcia Cetertich Bazzan
• Modalidade de apoio: Auxílio à Pesquisa - Regular