K. D. Reilly,
J. Barrett
J. Tarng
R. Hyatt
A Computerized Formal Means
To Reason About Components
In
Simulation Models And Environments
Part I: A Logic-Based Reasoning Methodology
Transactions of the Society for Computer Simulation, 12, 2, 101-178.
Abstract
The modeling and simulation enterprise is rapidly expanding due to an influx
of ideas and methods from artificial intelligence and simulation theory. This
paper proposes a blending of theory, a formal logic based reasoning system,
with AI-inspired symbolic programming methodology to develop an executable
formal reasoning system. The principal focus is on runnable specifications and
includes topics such as the formal system's roots, its scope and purpose, how
it works, and how it may evolve. With reliance on appendix material to relate
the work to others' and to define both the modeling and the environment
problems in a general purpose context, the paper addresses such topics as
procedural and non-procedural modes of reasoning, a dual reasoning mode to
provide a modeler-accommodating reasoning system, automatic and semi-automatic
translation from requirements and prior implementations to specifications and
from specifications to implementations. Logic and logic programming are joined
by a system and software scheme we've developed to promote symbolic computation
within a context of high level numerically strong languages and, equally
important, to support modeling life cycle activities, e.g., specifications
creation, testing, development and use. Much of the discussion motivates and is
motivated by material covered in part II of the paper. The ensemble of logic,
life cycle, and software development theories spanning both parts of the paper
constitutes a theory for a ``general purpose simulation environment" with an
executable formal reasoning capability as an integral part.
Key Words:
Logic-Based Reasoning in Simulation, Formal Methods, Relational Semantics, Logic
Programming, Automatically Generated Implementations Generating Specifications,
AI and Simulation, Combined Numerical-Symbolic Simulation, General Purpose
Simulation Environments.