A DSM-based Multi-paradigm Composable Modeling Method For Effectiveness Simulation

The combat effectiveness measurement of equipments lays a critical foundation for the demonstration, development and training of equipments, which can perform quantitative analysis on the ability of equipments to fulfill specified combat tasks under typical scenarios. Different sides of equipment system-of-systems(ESo S) and personnel are involved in the combat effectiveness measurement which needs to cope with complex counterwork process and a large amount of uncertain factors. Since the limitation of traditional analytical methods and the unavailability of physical experiments, the simulation based method, a.k.a., simulation based combat effectiveness measurement, is introduced into the combat effectiveness measurement of equipments.On the one side, combat effectiveness simulation(CES) should be performed under the condition of ESo S counterwork in several scenarios, thus there is a prominent requirement for composable modeling; on the other side, ESo S comprises multiple domains, so domain-specific modeling(DSM) should be conducted to raise the modeling abstraction level. To meet these two modeling requirements, this dissertation makes an integrative research on five problems of model design and implementation, and constructs a modeling framework from the aspect of simulation modeling lifecycle and studies its methods and key techniques. The contents and primary contributions to research include the following four parts:(1) A DSM based CES multi-paradigm composable modeling framework is proposed. Four kinds of current unified modeling methods can not support composable modeling and domain modeling sufficiently, thus a DSM-based CES MPM composable modeling framework is proposed, which uses a unified model framework and employs DSM to construct domain-specific composable modeling framework with the support of multiple formalisms, and its constitute modeling methods, modeling process and key techniques are generally described to provide methodological guide and technical support. This framework provides domain-specific modeling paradigms based on domain neutral modeling paradigms, raises the modeling abstraction level from formalism to problem domain, supports model composition of multi-domain models using multi-method, and enables rapid development of CES applications.(2) A SMP2 based CES architecture modeling method is proposed. Though the idea of model framework is common in the development process of simulation applications, informal and implicit methods are usually used to describe the model framework. This dissertation uses three modeling levels of MDA to formally and explicitly design and implement the CES architecture based on SMP2, which transforms domain ontology to SMP2 model framework, and generates executable models at platform specific model level using SMP2/C++ mapping. This method provides formal structural support for DSM at conceptual modeling level and model framework-based integration mechanism to compose multi-domain models and multi-formalism models.(3) A state-centered composable modeling method is based on the research foundation of multi-formalism CES behavioral modeling. Firstly, based on the analysis of equipment behavioral characteristics, a behavioral characteristics-oriented model design approach is proposed; secondly, a state-centered multi-formalism composable modeling method is researched; thirdly, the formalism-based model analysis method is discussed; finally, the automatic implementation and model integration are researched. The state-centered composable modeling method uses internal state-oriented formalisms as the center, and establishes the mapping and composition relationships of the state formalisms and other characteristics-oriented formalisms at meta-model level; thus it supports the appropriate usage of multiple formalisms to describe the behavioral characteristics from multiple aspects of the equipments.(4) A domain specific simulation modeling based composable modeling method is proposed based on relevant research on DSM, including the following steps: firstly, a ontology-structure-behavior metamodelling method is proposed based on DSM research fruits form both the software engineering community and M&S community, which comprehensively uses the external interface information of effectiveness simulation ontology, structural information from the sub-domain ontology and behavioral characteristics from the formalisms, and meta-models are built and used to generate domain specific modeling environment; secondly, model transformation techniques are used to transform domain specific simulation modeling language to formalisms to support model analysis; finally, the code generation based automatic implementation and model framework based integration of domains specific models are researched.Finally, the application of the modeling framework is researched based on the above-mentioned research foundation: firstly, multi-paradigm modeling(MPM) framework and its supporting methods are used to conduct composable and domain specific modeling on typical equipment in typical domains; Then the composable development of effectiveness simulation application is performed using case studies to validate the proposed modeling framework, methods and techniques.This dissertation proposes to customize domain specific modeling paradigms according to problem domain features based on formalisms, which raises the abstraction level of MPM from domain neutral formalism level to problem domain level, thus it can support composable modeling and domain modeling efficiently. This dissertation proposes a solution to perform DSM-based MPM composable simulation modeling under the constraint of the model framework, which are quite meaningful to simulation modeling research of other complex systems which also have multi-domain and composable modeling requirements.