A Manufacturing Information Considered Cell Formation Method Using A Machine-Operation-Part Matrix and A Hybrid Genetic Algorithm

• Main Authors: Hsieh, Chi-Kai, 謝智凱
• Other Authors: Yang, Feng-Cheng
• Format: Others
• Language: zh-TW
• Published: 2001
• Online Access: http://ndltd.ncl.edu.tw/handle/35845261024028481091

碩士 === 國立臺灣大學 === 工業工程學研究所 === 89 === The well-known two-dimensional machine-part relational matrix was widely used and resulted in impractical and rough cell formation. Previous research usually did not consider the manufacturing information and resulted in non-mature cell formation. This paper presents a cell formation method based on a three-dimensional machine-operation-part relational matrix. Hybrid genetic algorithm techniques are used and manufacturing information is included to generate practical manufacturing cells. In the presented method, hybrid genetic algorithms are adapted to model and solve the cell formation problems. In genetic coding, manufacturing information about parts demand and part routings is included. The objectives of the cell formation approach are to minimize inter-cell part movements and to lower load variations among cells. A prototype software system is developed on PC computers using Borland C++ Builder programming tools. Friendly user interfaces are provided for easily sample testing. Due to the fact that parameter settings and operator selections may significantly influence the computation results of genetic models, experiment design is conducted to test these variations and options. Two existing and published data samples and two generated ones are tested to verify the validity of the presented method. Results show that the method can generate useful and practical cell formation.