A machine-learning approach to the generation of assembly sequences

摘要

The authors describe an approach to acquiring heuristic knowledge automatically for generating assembly sequences efficiently. The approach induces the heuristic knowledge from training examples, based on ID3 algorithm. Using the heuristic knowledge, a method developed before can generate the assembly sequences efficiently. Additional training examples are made from the information on the assembly sequences generated by the method, and they are used to modify the heuristic knowledge. As the assembly sequence generation and the modification of the heuristic knowledge are executed more times, the heuristic knowledge is refined more and the assembly sequences can be generated more efficiently. An experiment is carried out to demonstrate the effectiveness of the approach.