曲げ工程の自動設計に対する遺伝的アルゴリズムの適用(機械要素,潤滑,工作,生産管理など)

概要

論文の詳細を見る
An automated design of bending process was carried out by using an artificial intelligence. This design which is to decide an optimum order of edges to be bent can be considered as a kind of traveling salesman problem. A genetic algorithm was applied for a solution, where a gene is an edge to be bent and chromosome is an order of gene. If an interference occurred due to an ill-considered order, a process cannot be carried out. The interference would to be known at the beginning of a design, while it often occurs at the end of a process. In order to design efficiently a practice of an inverse process design, which means not to decide a beginning order from a raw material to a product but to do a development order from a product to a raw material. Automatic design of two product profiles were carried out in the condition that two crossovers were examined and a scaling changed and it done successfully in the case of combination of two point and permutation crossover and gradual change of weight of a scaling. A fitness function was composed of five process factors which include a interference.
一般社団法人日本機械学会の論文
2002-07-25