Assembly sequencing through graph reasoning : graph grammar rules for assembly planning

Abstract

Assembly planning is difficult and tedious, but is necessary for complex products. This thesis presents a novel approach to automating assembly planning utilizing graph grammars. Computational geometric reasoning is used to produce a label rich graph from a CAD model. This graph is then modified by graph grammar rules to produce candidate assembly sequences which are run in conjunction with a tree search algorithm. An evaluation system then evaluates partial assembly sequences, which are used by the tree- search to find near-optimal assembly sequences.