Design Healing framework for automated code compliance

Automated Compliance Checking (ACC) techniques have advanced significantly, enabling designers to evaluate building designs against codes. However, architectural engineers have to improve the design by manually implementing the ACC results, which is laborious, iterative, and requires domain expertise. To address this challenge, this paper introduces a Design Healing framework that adapts the original design into code-compliant alternatives. By integrating design data and ACC results, the framework identifies critical non-compliant components through a graph-based topological algorithm and sensitivity analysis. A prior knowledge-informed design space exploration is conducted to find valid alternatives and quantify modifications using weighted Euclidean distances, allowing designers to select options closely aligned with the initial design. Multi-scenario experiments demonstrate the framework's effectiveness in resolving architectural spatial design violations. By automating post-checking design adaptation, the framework reduces manual revisions and provides an efficient tool for achieving compliance while accommodating varying design constraints. This paper establishes a basis for advancing designer-centered automated design correction methods based on ACC techniques.