Natural forming multi-agent systems (aka Swarms) have the ability to grow to enormous sizes without requiring any of the agents to oversee the entire system. The success of these systems comes from the fact that agents are simple and the interaction with the environment and neighboring agents is local in nature. In this paper we look at abstractions in the field of swarms and study their applicability in the context of coordination systems. In particular, we focus on the problematic issue of scalability of Linda systems. The purpose of this work is to look at abstractions yielded from observations of swarms and the way they are organized, and demonstrate how these abstractions may be used to implement a scalable tuple distribution mechanism in a Linda system -- to be named SwarmLinda.
Menezes, R., Tolksdorf, R. (2003). A new approach to scalable Linda-systems based on swarms (EXTENDED VERSION) (CS-2003-04). Melbourne, FL. Florida Institute of Technology