FIRST EXAMPLE: SIMULATING A ROBOTIC TANK
This example serves two purposes. First, it illustrates how hybrid dynamics can appear in engineering problems. The model has three main parts: the equations of motion, a model of the propulsion system, and a model of the computer. The first two are piecewise continuous with discontinuities caused by step changes in the motor voltage and the sticking friction of the rubber tracks. The third model is a prototypical example of a discrete-event system; the tank’s computer is modeled with an interruptible server and queue. The equations of motion, propulsion system, and computer are combined to form a complete model of the tank.
Second, this example illustrates the basic elements of a software architecture for large simulation programs. The simulation engine is responsible solely for calculating the dynamic behavior of the model; other functions (visualization and interactive controls, calculation of performance metrics, etc.) are delegated to other parts of the software. This approach is based on two patterns or principles: model–view–control and the experimental frame.
Model–view–control is a pattern widely used in the design of user interfaces (see, e.g., Refs. 47 and 101); the simulation engine and model are treated as a dynamic document and, with this perspective, the overarching design will probably be familiar to most software engineers. The experimental frame (as described, e.g., by Daum and Sargent )1 is a logical separation ...