Constraint-based approach for automatic spatial layout planning

Spatial layout planning generates two-dimensional layouts consisting of configuration of rectangles, and our new method represents the rectangles as an inconsistent sets of linear constraints. After that, we select the consistent sets from them by using constraint processing language. Because this type of problem is of the NP-complete class, we also introduce some redundant constraints such as qualitative relation and capacity. Redundant constraints generally slow a system, but with the partial (incremental) solver used for the´s type of problem, redundant constraints may be useful in reducing the search space. Our approach is now being applied to the layout of personal homes. In this paper, we use this example to explain our approach and we describe our floor planning system TG-FP (Tokyo Gas Floor Planner) implemented by CLP (Constraint Logic Programming) Language. TG-FP can assist an architect by providing a large variety of feasible plans. In TG-FP, the only information that must be provided by the user is the definition of the room. We also point out the advantages of our approach with CLP language: advantages such as a declarative paradigm for easy description and efficiency for avoiding combinatorial explosion