STRUCTURAL SIMULATION OF DISTRIBUTED
COMPUTATIONS USING CHRONOLOG(MC)

CHUCHANG LIU and MEHMET A. ORGUN
Department of Computing, Macquarie University
Sydney, NSW 2109, Australia
E-mail: fcliu, [email protected]

ABSTRACT

Chronolog(MC) is a logic programming language based on a linear-time temporal logic with multiple granularity of time. In this logic, all formulas can be clocked and are allowed to be defined on local clocks. Therefore the language is flexible and it is suitable for modeling and specifying those systems where the notion of dynamic change is central. This paper presents a knowledge-based modeling methodology for the specification of distributed computations with Chronolog(MC). The methodology combines trajectory and structural simulation methods together to describe a simulation system, and it develops a modeling formalism to represent the initial structural representation of a system as well as a method to use in determining the next representation of the structure. This methodology can be used for building a simulation environment for distributed computations, which provides an integration of modeling formalism, modeling environment, knowledge-base management and processing of the model in simulation.

1. Introduction

There has been a growing interest in knowledge-based simulation methods. There are a number of approaches in knowledge-based modeling and simulation tools, for instance, Gallaire5 proposed a language for modeling based on logic programming and object programming, MASCOT11, a simulation modeling and training environment, is a Prolog-based system, and SimKit17 built on top of knowledge engineering environment is an integrated set of general-purpose simulation and modeling tools. Knowledge-based simulation involves several model-related concepts such as modeling formalisms, modeling environments, model-base management and processing of models15. Since the use of models is essential in simulation, providing good simulation languages and implementation of environments to make integration of the above concepts are highly desirable.

Simulation requires models of a process to be simulated and it deals with processes evolving in time. Therefore, the timing properties of a system must be considered in simulation. There have been several simulation languages based on temporal logic such as SimTL18 and Simian3. SimTL incorporates a simulation language based on logic programming paradigm and a query language based on temporal logic, and it provides support for object-oriented programming. The designers of Simian especially consider the kinds of control structures that are necessary to accommodate the variety