The paper introduces the R*-tree, an efficient and robust access method for points and rectangles. The R-tree, a popular method for storing rectangles, optimizes the area of enclosing rectangles in inner nodes. However, this heuristic approach can lead to suboptimal performance. The authors designed the R*-tree by incorporating a combined optimization of area, margin, and overlap of each enclosing rectangle in the directory. Through extensive experiments, they found that the R*-tree outperforms existing R-tree variants, including Guttman's linear and quadratic R-trees and Greene's variant, in various types of queries and operations, such as map overlay. The R*-tree is particularly advantageous for supporting both point and spatial data efficiently and has a slightly higher implementation cost compared to other R-trees. The paper also discusses the design of the R*-tree, including its insertion and deletion algorithms, and presents performance comparisons with other R-tree variants, demonstrating the R*-tree's superior robustness and efficiency.The paper introduces the R*-tree, an efficient and robust access method for points and rectangles. The R-tree, a popular method for storing rectangles, optimizes the area of enclosing rectangles in inner nodes. However, this heuristic approach can lead to suboptimal performance. The authors designed the R*-tree by incorporating a combined optimization of area, margin, and overlap of each enclosing rectangle in the directory. Through extensive experiments, they found that the R*-tree outperforms existing R-tree variants, including Guttman's linear and quadratic R-trees and Greene's variant, in various types of queries and operations, such as map overlay. The R*-tree is particularly advantageous for supporting both point and spatial data efficiently and has a slightly higher implementation cost compared to other R-trees. The paper also discusses the design of the R*-tree, including its insertion and deletion algorithms, and presents performance comparisons with other R-tree variants, demonstrating the R*-tree's superior robustness and efficiency.