Simulation Modeling and Analysis
Simulation Modeling and Analysis
| Hi-index | 0.00 |
Expert-VSim is the intended product of a current research effort. It is a software system that provides a complete, intelligent environment for the construction and simulation of discrete event models. The initial stage consisted of the construction of a simulation environment called VSim [1] and the second involves the integration of an expert system.VSim provides a highly interactive environment. Included is a graphics interface which constitutes the main model construction tool. Network models are used to describe real-world systems through an entity-activity world view. Models can be built and tested incrementally, increasing the confidence of correctness. In addition, the multi level structure of VSim allows the user to define subnetworks as nodes within a larger network. A definition language complements the graphics interface and is used to describe the details of the model as well as to create user-defined nodes. A prototype for the VSim stage has been completed and thoroughly tested.The design of the second stage is in active development and near completion. This design includes the creation of a simulation and statistics expert system to be integrated with VSim to obtain the final software product: Expert-VSim. This expert system is rule-based implemented in Prolog.The presence of such an expert system is of significant importance in the modeling and simulation process. It provides database and inference capabilities that assist the user in:the selection of probability distributions for input variables using common heuristics and procedures [2] to choose from the known theoretical distributions and estimate their parameters (with or without that availability of data), or to allow the user to fit a distribution to experimental data; the analysis of the simulation results, the construction of confidence intervals for the observed variables, and the comparison of simulation runs;the construction of the model itself by providing an on-line database containing the descriptions of the standard and user-defined node types together with the ability to traverse and inspect the multi-level structure of the model;the extension of the expert system itself by allowing the user to add to the set of rules and facts through the Prolog interface;the independent use of the statistical analysis functions and procedures available to analyze and process data;and the management of experimental and simulation data through the database capabilities of the system.