We gratefully acknowledge support from
the Simons Foundation and member institutions.
Full-text links:

Download:

Current browse context:

math.NA

Change to browse by:

References & Citations

Bookmark

(what is this?)
CiteULike logo BibSonomy logo Mendeley logo del.icio.us logo Digg logo Reddit logo

Mathematics > Numerical Analysis

Title: Continuous Optimization of Adaptive Quadtree Structures

Authors: Jun Wu
Abstract: We present a novel continuous optimization method to the discrete problem of quadtree optimization. The optimization aims at achieving a quadtree structure with the highest mechanical stiffness, where the edges in the quadtree are interpreted as structural elements carrying mechanical loads. We formulate quadtree optimization as a continuous material distribution problem. The discrete design variables (i.e., to refine or not to refine) are replaced by continuous variables on multiple levels in the quadtree hierarchy. In discrete quadtree optimization, a cell is only eligible for refinement if its parent cell has been refined. We propose a continuous analogue to this dependency for continuous multi-level design variables, and integrate it in the iterative optimization process. Our results show that the continuously optimized quadtree structures perform much stiffer than uniform patterns and the heuristically optimized counterparts. We demonstrate the use of adaptive structures as lightweight infill for 3D printed parts, where uniform geometric patterns have been typically used in practice.
Comments: Solid and Physical Modeling - SPM 2018
Subjects: Numerical Analysis (math.NA); Computational Engineering, Finance, and Science (cs.CE); Graphics (cs.GR)
Journal reference: Computer-Aided Design 102 (2018) 72-82
DOI: 10.1016/j.cad.2018.04.008
Cite as: arXiv:1712.04523 [math.NA]
  (or arXiv:1712.04523v2 [math.NA] for this version)

Submission history

From: Jun Wu [view email]
[v1] Tue, 12 Dec 2017 21:13:36 GMT (1905kb,D)
[v2] Tue, 22 May 2018 08:56:03 GMT (6232kb,D)

Link back to: arXiv, form interface, contact.