Authors:
Torsten Spieldenner
1
;
2
and
Melvin Chelli
1
Affiliations:
1
German Research Center for Artificial Intelligence (DFKI), Saarland Informatics Campus D3 2, 66123 Saarbrücken, Germany
;
2
Saarbrücken Graduate School of Computer Science, Campus E1 3, 66123 Saarbrücken, Germany
Keyword(s):
Stigmergy, Stigmergic Medium, Coordination, Multi-Agent Systems, Linked Data, Semantic Web, Nature-inspired Algorithm.
Abstract:
Algorithms inspired by nature have gained focus in research as a solution to classic coordination and optimization problems. A certain type of these algorithms employs principles of stigmergy: in stigmergic systems, coordination arises from agents leaving traces of their actions in the environment, or medium, that they work on. Other agents instinctively adapt their behavior based on the traces, by which, in the end, the fulfillment of a higher goal emerges from elementary actions of many, rather than thorough planning of complex actions of a few. Despite the perceivable uptake of stigmergic algorithms for coordination in various domains, a common clear understanding of a suitable digital stigmergic medium is lacking. It should however be assumed that a well-defined, properly modelled, and technically sound digital medium provides a crucial basis for correct, efficient and transferable stigmergic algorithms. In this paper, we motivate read-write Linked Data as generic medium for dece
ntralized stigmergic coordination algorithms. We show how Linked Data fulfills a set of core requirements that we derived for stigmergic media from relevant literature, provide an application example from the domain of digital manufacturing, and finally provide a working example algorithm for stigmergic decentralized coordination.
(More)