Observations on real-time simulation design and experimentation

Modeling and simulation provides a convenient way for evaluating operability of designed systems (e.g., Network-on-Chip) in their desired target environments. Where applicable, real-time modeling with real-time simulation extends logicaltime M&S capabilities by evaluating the degree to which system requirements can be satisfied in actual operational scenarios. Simulations can be directly observed and evaluated in realistic settings. This brings about a well-known challenge - simulation experimentations are never without computational cost. This leads to real-time simulation (adversely) affecting simulation results, sometimes drastically, if not carefully designed for. In this paper, in view of finite computational resources for executing simulations in real-time, new observations on data collection and evaluations are described for the Action-Level Real-Time DEVS (ALRT-DEVS) simulator. Simple and complex example real-time models are developed in selected simulators, experiments with varying degrees of data collection volume are conducted, and results are discussed.