Abstract: In response to the problems of low frame rate, high memory consumption, and average CPU consumption in the scheduling process of large-scale BIM scenes, an adaptive scheduling method for large-scale BIM scenes based on multidimensional perception regulation is proposed. Divide large-scale BIM scene spatial data into DEM data, building data, and texture data, and implement spatial data segmentation separately. Combining the pyramid structure with the quadtree data indexing mechanism to implement spatial data indexing for segmented data blocks, clearly expressing the internal structural relationships of the data blocks. Design a multidimensional perception adaptive scheduling algorithm based on a multidimensional perception regulation model and Markov decision-making, and implement adaptive scheduling for large-scale BIM scenarios based on the designed index mechanism. The test results show that the design method can achieve large-scale BIM scene scheduling. When the data size reaches 1000GB, the frame rate can be maintained at 39fps, the average CPU consumption is less than 1800MB, and the memory consumption is relatively low.

Key words: spatial data segmentation, multi dimensional perception regulation model, Markov decision-making, large scale BIM scenarios, adaptive scheduling

摘要： 针对大规模BIM场景调度过程中,存在帧率偏低、内存消耗与CPU平均消耗较高的问题,文中提出基于多维感知调控设计大规模BIM场景自适应调度方法。将大规模BIM场景空间数据分为DEM数据、建筑数据、纹理数据,分别实施空间数据分割。结合金字塔结构与四叉树的数据索引机制实施分割数据块的空间数据索引,明确表达数据块的内部结构关系。基于多维感知调控模型与马尔可夫决策设计多维感知自适应调度算法,基于设计的索引机制实现大规模BIM场景的自适应调度。测试结果表明,设计方法能够实现极大规模的BIM场景调度,在调度数据规模达到1000GB时,其帧率可以保持在39fps,CPU平均消耗低于1800MB,内存消耗较低。

关键词: 空间数据分割, 多维感知调控模型, 马尔可夫决策, 大规模BIM场景, 自适应调度