ONTOLOGY-BASED ADAPTIVE OBJECT MODEL FOR SIMULATION PHYSICAL PARAMETER MANAGEMENT

Simulation is widely used. It requires a great deal of various physical parameters, some of which are existing, some of which demand to be measured or tested. For a simulation parameter management system, the parameters should be queried easily and new type of parameters, rather than new parameters of the existing type, could be added dynamically. This paper presents a plug-and-play system for simulation physical parameter management based on ontology and AOM (Adaptive Object Model). It consists of a three-layer model: meta model layer, domain model and data instance layer. An interpreter engine is built according to the meta model layer. Domain expert, with little programming experience, could make or modify domain model by instantiating the meta model through Protege. According to the domain model, interpreter engine could generate templates of both excel and database. The former acts as an importer for data, and the latter would be the model of a particular data record, taking advantage of non-predefined-schema NoSQL database. Thus, the system could be adaptive for any domain by modifying the domain model. Separated from the program itself, the domain model exists as an independent configuration file, which means this model could be edited with immediate effect during run-time. By collecting all the data of the same template, a statistic would be calculated. Once a new data is imported,the statistic would evolve with it. Applied in the development of the physical parameter management system for the flexible paper-like object in an ATM manufacturer, the system is demonstrated to be effective.