A FRAMEWORK FOR THE DEVELOPMENT OF SIMULATION SYSTEMS (COMPUTATION THEORY, COMPILING EXPERT)

A framework for the development of computer simulation systems based on interpreters is to be developed. These interpreters are to be developed independently of any preconceived structure of the simulator and are to be combined dynamically by the modeler. The foundation for the development of this framework is computability theory and denotational semantics.;An interpreter based on these new concepts is also described and examples of its use are given. This interpreter combines many ideas from operating systems and macro processors. The structure of the algorithm and the computational structure model indicates a different machine paradigm than those currently in use. The salient point is that the interpreter uses the computational structure paradigm to construct strings which are then queued for interpretation by an appropriate computational structure.;Two technical results of interest are developed. The first result is a new canonical form for programs, called the state-transition form. This form provides for only one block type which has one entrance but multiple exits. Using this form, we prove that the runtime trace of a program is a regular expression. Secondly, we give a nonconstructive proof that recursive functions have a set of basis functions which are based on pushdown automata.;Future directions are discussed, particularly in extending the algebraic aspects of computational structures. We discuss mathematical modeling system vis-a-vis expert systems and conclude that the two types are different in their goals. Finally, some new views of computation are presented and applications to computational mathematics are discussed.;A unified approach to syntax and semantics in interpreters is presented. A new concept, the typed Herbrand universe, is developed. This concept allows a semantic and pragmatic structure to be imposed on a (formal) language. Using the typed Herbrand universe, we formalize the notion of evaluation functions. The evaluation functions must have a property of forming a basis; this second new concept which we term the computational system property. The third new concept developed is the computational structure. A computational structure is a morphism from one formal language to another. Here, the term morphism is used in a general way to mean a function between formal languages which preserves some properties. The computational structure unifies ideas concerning the syntax, semantics, and pragmatics of language.