Dynamic Constrained Objects for Vehicular Network Modeling

Abstract

We present a paradigm called dynamic constrained objects for a declarative approach to modeling complex systems. In the basic paradigm of constrained objects, the structure of a complex system is specified through objects (as in object-oriented languages), while the behavior of a complex system is specified declaratively through constraints (as in constraint languages). The emergent behavior of such a complex system is deduced through a process of constraint satisfaction. Our focus in this paper is on systems whose states change with time. Such time-varying behaviors are fundamental in many domains, especially in mission and safety-critical applications. We present an extension of constrained objects with special metric temporal operators over time-series data, and we discuss their properties. We refer to the resulting paradigm as dynamic constrained objects and we illustrate their use for vehicular network modeling. Here, the network of roads and the roadside infrastructure are specified through objects, and the movement of vehicles and associated safety and liveness conditions are specified through time-series variables and metric temporal operators. The paper presents a language called DCOB, for dynamic constrained objects, and examples of its use for vehicular network modeling.