Research On The Common Framework For Execution-data Processing Of Federated Simulation

Federated simulation is one of recent trends of advanced types of simulation, which can be used to explore complex problems. It supports various operation modes and expands the exploration space, but also leads to increasing computational complexity. So, the more powerful capability of data processing is urgently needed. Such a capability should generally contain two aspects: technique and management. This dissertation explored specifically the technique issue which was more foundational, and hoped to construct a common framework for federated simulation's execution-data processing, which should be reusable, composable, extensible, general, roboticized, high-efficiency and interoperable. The motivation was the desire to build an infrastructure for the data processing, which was similar to the industrialized product line that could utilize interchangeable parts, production assembly line and robots to automate the manufacture process. So it could not only meet various data-processing requirements, but also make related work become long-term investment. Such a framework should overcome the shortcoming of traditional solutions and have advantages as follows: not limited to any certain simulation technology standard, focus from engineering single system to engineering families of systems, more open system architecture, roboticized process of code generating and components assembling /deploying, self-evolvable system elements, and supporting dynamic evaluation or reconfiguration during execution, and so on.Therefore, this dissertation defined a technical framework dedicated for federated simulation's execution-data processing, parsed the concept of common framework, and expounded on the design approach and the integrated implemented process. It shifted from the traditional practice of manually developing"one solution for one application"to a new approach to resolve this kind of problems. Around the theme"designing the technical framework", some crucial design principles were emphasized that included the design of top-level architecture oriented systems'family and data interoperability capability, and the two complete development cycles: one for designing and implementing the reusable domain components that to develop for reuse, the other for using the components to produce concrete system for specific applications that to develop with reuse. In practice, it was tried to simultaneously ensure the framework's functional requirements and non-functional requirements via suitable software architecture design and development technique.The dissertation started with exploring data activities in the life cycle of federated Modeling&Simulation and analyzing requirements derived from them, put forward the way and the principle of constructing common framework based on the definite mission space of federated simulation's execution-data processing, and then investigated the problems related to architecture design such as how best to partition a system, how components identify and communicate with each other, how information communicate, how elements of the system evolve independently, and so on. Based on these work, the dissertation made an in-depth study aim at the two pivotal issues that one was the constructing of the Config&Control Platform, the other was the implement of the capability of data interoperability and persistent storage. To meet the demand for the roboticized software product line, it were established that the hierarchical framework model and corresponding component system, the mechanism of config, managing, operating and controlling the tasks process, and the mechanism of supporting data interoperability and persistent storage. Finally, the dissertation summarized the inbeing of various capabilities of common framework as"eight kinds of data exchange relationship among relative entities, which were around the operating of the core object (execution-data)", and verified the validity of the design and technique via the cases such as HLA data collection and data playback. As the finis of this dissertation, the future of common framework and the farther improvements needed were discussed.