English (United Kingdom)

https://curated-unify.zendy.io/wp-json/zendy-region/v1/featured_content/oa?rat=en

https://curated-unify.zendy.io/wp-json/zendy-region/v1/highlighted_journal/

Global: Zendy Plus annual

Presents the access of premium content as premium feature

Premium Content

Presents the keyphrase highlighting as premium feature

Keyphrase Highlighting

Presents the summarisation as premium feature

Summarisation

Insights

Presents the pdf analysis as premium feature

PDF Analysis

Presents the zaia usage as premium feature

ZAIA

Global: Zendy Plus monthly

Global: Zendy Tools annual

Global: Zendy Tools monthly

Global: Zendy Free Plan

We study the problem of learning online packing skills for irregular 3Dshapes, which is arguably the most challenging setting of bin packing problems.The goal is to consecutively move a sequence of 3D objects with arbitraryshapes into a designated container with only partial observations of the objectsequence. Meanwhile, we take physical realizability into account, involvingphysics dynamics and constraints of a placement. The packing policy shouldunderstand the 3D geometry of the object to be packed and make effectivedecisions to accommodate it in the container in a physically realizable way. Wepropose a Reinforcement Learning (RL) pipeline to learn the policy. The complexirregular geometry and imperfect object placement together lead to hugesolution space. Direct training in such space is prohibitively data intensive.We instead propose a theoretically-provable method for candidate actiongeneration to reduce the action space of RL and the learning burden. Aparameterized policy is then learned to select the best placement from thecandidates. Equipped with an efficient method of asynchronous RL accelerationand a data preparation process of simulation-ready training sequences, a maturepacking policy can be trained in a physics-based environment within 48 hours.Through extensive evaluation on a variety of real-life shape datasets andcomparisons with state-of-the-art baselines, we demonstrate that our methodoutperforms the best-performing baseline on all datasets by at least 12.8% interms of packing utility.

Learning Physically Realizable Skills for Online Packing of General 3D Shapes