Optimal Index Partitioning of Main-Memory Based TPR*-Tree for Real-Time Tactical Moving Objects

The main-memory based TPR*-tree for real-time tactical moving objects is most effective for real-time prediction query processing. The target objects collected from radar are mixed with high-speed objects of aviation in addition to the low-speed objects in the sea, and there is a problem that the overlap area of the MBR increases in the main-memory based TPR*-tree. This paper proposes an optimal index partitioning to minimize the overlap of MBR in main-memory based TPR*-tree. An optimization partitioning algorithm is proposed in which both the cost of space-time prediction query processing is minimized and the cost of combining processing of divided indexes is minimized.