Modelling large scale enterprises : A distributed simulation approach.

Distributed simulation provides an alternative solution when today's highly complicated systems including manufacturing are to be simulated. Complexities involved in implementation, the need for more expertise, high development cost and long implementation time etc. along with a lack of guidelines for developing distributed simulation, and the complexity of tools and techniques used to implement schemes, resulted in the lack of acceptance for distributed simulation among the general simulation community. In order to address some of these issues, a new approach is proposed for modelling and simulating large scale enterprises using distributed simulation. The proposed approach which includes a comprehensive methodology for distributed enterprise simulation, developed by combining activities required for traditional sequential simulation with additional activities required for distributed simulation. The thesis elaborates the additional activities required for distributed simulation in different chapters with simplified approaches for carrying out these activities. These include an approach to decide the appropriate simulation strategy (SimSS process), a simplified approach to modelling and model partitioning, a synchronization mechanism that approximately synchronizes the distributed enterprise simulation and an approach for developing distributed simulation using tools and technologies which are popular, well accepted and also cost effective. The differences between the traditional distributed simulation approaches and the proposed methodology include: partitioning of the conceptual model into logical processes before transforming them into computer simulation models, use of commercial simulation software to implement the distributed simulation, and use of cost effective and popular middleware and programming languages. Illustration of the proposed approaches focuses on distributed manufacturing applications.