Cooperative Multi-agent Learning in a Large Dynamic Environment

Hierarchical Reinforcement Learning (HRL) enables autonomous decomposition of challenging long-horizon decision-making tasks into simpler subtasks. During the past years, the landscape of HRL research has grown profoundly, resulting in copious approaches. A comprehensive overview of this vast landscape is necessary to study HRL in an organized manner. We provide a survey of the diverse HRL approaches concerning the challenges of learning hierarchical policies, subtask discovery, transfer learning, and multi-agent learning using HRL. The survey is presented according to a novel taxonomy of the approaches. Based on the survey, a set of important open problems is proposed to motivate the future research in HRL. Furthermore, we outline a few suitable task domains for evaluating the HRL approaches and a few interesting examples of the practical applications of HRL in the Supplementary Material.

Download Full-text

The possible and the impossible in multi-agent learning

Artificial Intelligence ◽

10.1016/j.artint.2006.10.015 ◽

2007 ◽

Vol 171 (7) ◽

pp. 429-433 ◽

Cited By ~ 16

Author(s):

H. Peyton Young

Keyword(s):

Agent Learning ◽

Multi Agent

Download Full-text

AGENT-BASED SYSTEM ARCHITECTURE FOR DYNAMIC AND OPEN ENVIRONMENTS

International Journal of Information Technology & Decision Making ◽

10.1142/s0219622003000537 ◽

2003 ◽

Vol 02 (01) ◽

pp. 105-133 ◽

Cited By ~ 23

Author(s):

ELHADI SHAKSHUKI ◽

HAMADA GHENNIWA ◽

MAHAMED KAMEL

Keyword(s):

System Architecture ◽

Rapid Growth ◽

Dynamic Environment ◽

Information Gathering ◽

Information Resources ◽

Architecture Design ◽

Agent Based ◽

Interface Agents ◽

Multi Agent ◽

Computing Environments

The rapid growth of the network-centered (Internet and Intranet) computing environments requires new architectures for information gathering systems. Typically, in these environments, the information resources are dynamic, heterogeneous and distributed. In addition, these computing environments are open, where information resources may be connected or disconnected at any time. This paper presents an architecture for a multi-agent information gathering system. The architecture includes three types of agents: interface, broker and resource agents. The interface agents interact with the users to fulfill their interests and preferences. The resource agents access and capture the content of the information resources. The broker agents facilitate cooperation among the information and the resource agents to achieve their desired goals. This paper provides the agents' architecture, design and implementations that enable them to cooperate, coordinate and communicate with each other to gather information in an open and dynamic environment.

Download Full-text