Automatic assembly sequences generation by pattern matching

The problem of generating all the feasible assembly sequences of a set of n parts to construct a mechanical object is formulated as a state-constrained travelling salesman problem. In generating all the assembly sequences presented, the precedence-logic relations among parts to assemble an object are transformed into a pattern-matching problem. The concept of the pattern-matching algorithm is to match liaisons or parts with one of the answers to obtain the currently last assembly operation. Equivalency between the proposed knowledge acquisition and general AND/OR precedence nominal precedence forms is also discussed. The algorithms presented are able to reduce the number of questions asked in generating the assembly sequences