Towards Automatic Blocking of Shapes using Evolutionary Algorithm

Evocube: A Genetic Labelling Framework for Polycube‐Maps

“…When selecting an element from the archive, we use a stochastic model favouring higher ranked solutions, similarly to Lim et al. [LYZ*20]. For an archive containing ranked solutions from ℓ 1 to $$\ell _n$$, the probability of randomly picking $$\ell _i$$ is as follows: $$\begin{equation*} P(\ell _i) = \frac{n - i + 1}{1 + \cdots + n} \end{equation*}$$…”

“…Machine learning and mesh generation . A variety of geometry and mesh generation problems have seen great advancements in the past decade via machine learning techniques [XZCOC12, MTP*15, LYZ*20, DLLK21]. Marcias et al.…”

“…Lim et al. [LYZ*20] designed an evolutionary algorithm that performs automatic blocking of a 2D manifold. They define a set of simple genetic operators on a set of points from which they robustly extract a quad layout.…”

An evolutional algorithm for automatic 2D layer segmentation in laser-aided additive manufacturing

“…Nonetheless, to execute the generated toolpath in actual LAAM, the deposition width needs to be altered continuously, which poses high operation cost. In addition, as pointed out by Lim et al, the medial axis-based approaches may lead to non-optimal configurations for regions of concavities [12]. Recently, Zhai et al decomposed the 2D layer of a porous model into a set of topologically simple shapes via a generalized Voronoi diagram, and then generated a continuous toolpath [13].…”

“…The entire process could be extremely time-consuming, thus severely hindering their application in LAAM. Another piece of related work is presented by Lim et al towards automatic blocking of 2D shapes [12]. An evolutionary algorithm was employed to automatically segment the input shape into quadrilateral meshes for computational fluid dynamics (CFD) simulation.…”