PICO: Procedural Iterative Constrained Optimizer for Geometric Modeling.

Area of Science:

• Computer Graphics
• Computational Geometry
• Design Automation

Background:

• Procedural modeling offers powerful generative capabilities but often lacks user control and guidance.
• Existing methods require complex rule-writing, limiting accessibility and iterative design exploration.

Purpose of the Study:

• Introduce PICO (Procedural Iterative Constrained Optimizer) and PICO-Graph, a novel procedural modeling framework.
• Enable intuitive user guidance and constraint-based design exploration for complex geometries.
• Facilitate rapid ideation and prototyping in design workflows.

Main Methods:

• PICO utilizes a PICO-Graph, a directed cyclic graph of geometry-generating operations and axioms.
• User-defined constraints (e.g., support, symmetry, motion) are integrated into an optimization framework.
• The system generates and optimizes geometric models iteratively, providing continuous user feedback and interactive control.

Main Results:

• PICO enables fast generation of complex and varied geometries through its graph-based approach.
• Efficient optimization allows for rapid exploration of design variants that satisfy specified constraints.
• Demonstrated applications include generating chairs, 3D printing supports, spinning objects, and procedural terrains.

Conclusions:

• PICO offers a powerful, constraint-driven approach to procedural modeling and generative design.
• The framework significantly enhances user controllability and reduces the need for explicit procedural rule definition.
• PICO is well-suited for early-stage design, rapid ideation, and prototyping applications.