| Title |
Similarity Evaluation of Construction Site Layout Plans Using Graph Edit Distance |
| Authors |
Seung-Min Lee ; Jongwoo Cho ; Saruul Ishdorj ; Ho-Yong Lee ; Tae Wan Kim |
| DOI |
https://dx.doi.org/10.6106/KJCEM.2026.27.5.060 |
| Keywords |
Construction Site Layout Planning; Graph Edit Distance; Similarity Evaluation |
| Abstract |
To effectively retrieve similar cases of construction site layout plans, a quantitative evaluation criterion for objectively measuring structural similarity between layouts is required. Existing pixel-based methods such as Intersection over Union (IoU) have limitations in that they fail to adequately capture the topological characteristics of spatial structures. To address this, the present study proposes a method that models the spatial connectivity structure as an attributed graph and quantifies structural similarity between layout plans using Graph Edit Distance (GED). Void spaces are decomposed based on the largest inscribed circle, and each spatial unit is represented as a node with a three-dimensional attribute vector comprising normalized center coordinates and radius. Edges are defined by two spatial relationships: a contact relationship representing adjacency between neighboring units, and a mutual visibility relationship indicating the possibility of unobstructed straight-line passage. GED is computed as the minimum cumulative cost of node and edge insertion, deletion, and substitution operations, serving as a quantitative similarity measure that comprehensively reflects differences in the scale, position, structural relationships, and creation and elimination of void spaces. The validity of the proposed method is verified through systematic experiments across four types of layout modifications building deletion, shape modification, new building addition, and site boundary change demonstrating its potential as a case retrieval support tool grounded in structural similarity, beyond mere visual resemblance. |