Modelling environments for distributed simulation

  • Authors:
  • Michael Lees;Brian Logan;Rob Minson;Ton Oguara;Georgios Theodoropoulos

  • Affiliations:
  • School of Computer Science and IT, University of Nottingham, UK;School of Computer Science and IT, University of Nottingham, UK;School of Computer Science, University of Birmingham, UK;School of Computer Science, University of Birmingham, UK;School of Computer Science, University of Birmingham, UK

  • Venue:
  • E4MAS'04 Proceedings of the First international conference on Environments for Multi-Agent Systems
  • Year:
  • 2004

Quantified Score

Hi-index 0.01

Visualization

Abstract

Decentralised, event-driven distributed simulation is particularly suitable for modelling systems with inherent asynchronous parallelism, such as agent-based systems. However the efficient simulation of multi-agent systems presents particular challenges which are not addressed by standard parallel discrete event simulation (PDES) models and techniques. PDES approaches based on the logical process paradigm assume a fixed decomposition into processes, each of which maintains its own portion of the state of the simulation. The interaction between the processes is fixed in advance and does not change during the simulation. In contrast, simulations of MAS typically have a large shared state, the agents' environment, which is only loosely associated with any particular process. In this paper, we present a model of the shared state of a distributed MAS simulation of situated agents. We consider the problems of efficient sensing, parallel actions and action conflicts, and present preliminary work on an approach to the simulation of the environment which addresses these issues.