Stigmergy in Multi Agent Reinforcement Learning

8 years 8 months ago
Stigmergy in Multi Agent Reinforcement Learning
In this paper, we describe how certain aspects of the biological phenomena of stigmergy can be imported into multiagent reinforcement learning (MARL), with the purpose of better enabling coordination of agent actions and speeding up learning. In particular, we detail how these stigmergic aspects can be used to define an inter-agent communication framework. 1 Stigmergy The term stigmergy describes a kind of cooperative phenomena that emerges in a group of simple animals, like ants. For instance, when in a colony of ants, a mass of dead ants is strewn all over it, the live ants are able to make progressively lesser and larger heaps of the carcasses until finally they pile them onto just one big heap, which is the ideal solution to the problem of ridding the colony of obstacles [2]. The stigmergic pull here is directly proportional to the amassed dead ants in any given spot. An ant is more likely to place a dead ant in a spot where there is already a mass of dead-ants rather than in a sp...
Raghav Aras, Alain Dutech, François Charpil
Added 30 Oct 2010
Updated 30 Oct 2010
Type Conference
Year 2004
Where HIS
Authors Raghav Aras, Alain Dutech, François Charpillet
Comments (0)