Dynamic grid-based method research in data distributed management

In distributed interactive simulation based on High Level Architecture (HLA), the aim of data distributed management (DDM) can efficiently reduce network traffic and save system resources. With the enlarging of distributed interactive simulation system, the simulation system performs poorly using the past DDM algorithm. In order to resolve the problem that distributed interactive simulation based on HLA costs too much simulation time and has a low data filter rate using past data distributed management algorithm. A dynamic grid-based algorithm in data distributed management is discussed in this paper thoroughly and develops a DDM model that uses dynamic grids for matching the publishing/subscribing regions and for data filtering. Further analyzing the past static and dynamic algorithms of data distributed management and theory proof of the relationship between the grid-cell and the update time, data filer, receiving time and queue time by simulation and experiment, the formula of the algorithm is done. The theory results point out that grid-cell is key element to impact data filter and time cost during the running of simulation system. Comparison researched of three methods, such as static grid-based, dynamic grid-based and linear programming is done in a simple petrochemical system to resolve the problem of products optimization and capability equilibrium. Conclusions show that the algorithm improves the rate of data filter and reduces the simulation time cost, and verifies the availability of the algorithm.