Distributed Simulation in Industry Defended on Friday, 21 October 2005

While distributed simulation is widely accepted and applied in defence, it has not gathered ground yet in industry. In this thesis we investigate the reasons behind this phenomenon by surveying the expectation of industry with respect to distributed simulation solutions. Simulation models in industry are mainly designed and developed in commercial-off-the-shelf (COTS) simulation packages. The existing distributed simulation architectures in defence, however, do not focus on coupling models created in COTS simulation packages. Therefore, in order to motivate the industrial community to easily accept and use distributed simulation, one should strive to couple models built in these packages. Further, coupling these models should be possible without needing too much extra effort from modellers. In this thesis, based on a survey with experts in domain, we propose a list of requirements for designing and developing distributed simulation architectures that would encourage the industrial community to accept and apply distributed simulation. Furthermore, we present a lightweight distributed simulation architecture which has been successfully applied in two industrial projects, and satisfies to a large extent the proposed requirements.

Keywords

Distributed simulation, Simulation in industry, Commercial-Off-The-Shelf (COTS) simulation packages, Distributed modelling of complex systems, Reusability, Cost-benefit ratio of distributed simulation in industry, Interview with simulation experts, Requirements for distributed simulation architectures, The High Level Architecture, The FAMAS Simulation Backbone Architecture, industrie, software