Showing papers by "Bernard P. Zeigler published in 1986"

Multifaceted, multiparadigm modeling perspectives: tools for the 90's

Abstract: Multiplicities of perspective inherent in modelling and simulation methodology are enumerated and rationales given for their existence. Characteristics of futuristic simulation environments which support flexible adoption of multiple perspectives are outlined. Finally, we discuss the construction of models which simultaneously embody differing perspectives. Advances in modelling methodology along these lines will constitute a quantum leap in tool sophistication which can greatly extend the domain of simulation application.

Discrete-event simulation: When prototyping of large, complex systems is impossible, this technique, based on known behaviors of interacting elements, is useful

Abstract: The authors describe how discrete-event simulation is used. The development of an abstract model of how the system to be simulated works, which is the first step in discrete-event simulation, is described. The system model itself can range from a basic conceptual type, such as a paper study of how the parts of a system are put together, to a precise set of mathematical relationships describing the system's behavior. Once the model has been specified and validated, a language-specific model and a corresponding simulation program are developed. The utilization of artificial intelligence techniques is examined, and the outlook for discrete-event simulation is considered.

Knowledge-based design of LANs using system entity structure concepts

Abstract: The paper illustrates an application of knowledge based system design concepts to design of local area networks. A system entity structure representation of a local area network (LAN) is developed. This representation unifies a variety of possibilities for LAN design architecture. In a design session, the LAN design domain is subsequently restricted by pruning the entity structure with respect to network design objectives. A LAN model is then synthesized using a production rule scheme. The benefit of the knowledge-based LAN design is briefly discussed.

Hierarchical modular modeling/knowledge representation

Abstract: This tutorial will emphasize concepts and methodology and will relate them to languages and software environments which are becoming available to support these concepts. We will show how high level specification of discrete event models with hierarchical and modular properties is crucial to the sound integration of knowledge representation approaches of artificial intelligence. We will discuss examples of hierarchical modular models exhibiting self-modifying structure capabilities and show how they may be implemented in conventional, as well as object-oriented symbolic languages. Finally, structuring of model bases for simulation environments will be presented and example tools illustrated.Specifically, we shall discuss the following topics:High level specification of discrete event models with hierarchical and modular properties using pseudo-code formalism.Implementation of such specifications in procedural languages: How to express such pseudo-code in conventional languages such as SIMSCRIPT II.5 as well as in LISP-based languages.Examples of such hierarchical modular models exhibiting self-modifying structure capabilities and embedded artificial intelligence.Model/Knowledge Base Design: Organizing models using the system entity structure to constitute a reusable knowledge base.Background for this tutorial may be found in (Zeigler 1984, 1985).