分析仿真是以复杂问题分析为目的的一类仿真活动,在国防及国民经济各领域的应用日益广泛。随着仿真应用的不断深入,如何提高分析仿真的运行效率成为急需解决的问题。讨论了分析仿真的计算特点,指出了当前国内采用HLA体制、基于分布式网络解决方案存在的不足,分析了分析仿真的潜在并行性及面临的挑战;从国外分析仿真应用发展趋势以及国内实际情况出发,提出了基于高性能并行计算的分析仿真系统层次式体系架构,该方案可充分挖掘分析仿真的并行性,能有效支持分析仿真应用的运行和开发;同时该方案采用层次式架构,各层次内部功能相对独立,使得整个系统更易于维护和升级。
Analytic Simulation is an effective approach to study and analyze complex system. As a peer methodology to experiment and theory, it is used more and more widely in the area of defense and economy. With the in-depth development of the applications of analytic simulation, to shorten the time of simulation run has become an urgent job. The characteristics of analytic simulations were discussed, the shortages of current I-ILA-based solution using distributed network were pointed out, and the latent parallelisms of analytic simulation were analyzed. A layered analytic simulation architecture based on parallel processing was put forward. This solution can exploit the parallelism of the analytic simulation applications and support them well, while the layered architecture forms a clear separation between the functionality of each layer, resulting in easier to maintain and upgrade.