Computational and Simulation Models of the Control System on Modelica

Abstract

When modeling network protocols, the choice of a model approach and a software implementation tool is a problem. The specificity of this subject area is that for the description of protocols usually the discrete-event approach is used. However, the discrete model approach has several disadvantages. It is poorly scalable, not well suited for describing dynamic systems. As an alternative to the discrete approach, a continuous approach is usually considered. But when modeling discrete events, continuous description becomes unnecessarily complicated and heavy. Events take the form of some restrictions on the continuous system, which are often not explicitly included in the continuous model, but have the form of additional semantic descriptions. The authors propose to use a hybrid (continuous-discrete) approach when modeling such systems. In the framework of the hybrid approach, the discrete system is recorded in a continuous form, and the events take the form of discrete transitions inherent in the approach. In addition, if it is based on the description of events, a simulation model can be obtained on the basis of a hybrid approach. This paper demonstrates the use of a hybrid approach to describe systems with control by the example of the interaction of the TCP protocol and the RED algorithm. The simplicity of creating both computational and simulation models of the system is demonstrated. The Modelica language is used as the implementation language.