Research on the Automatic Optimization Algorithm on System Design using the System Definition

• Published in: Journal of Smart Processing Vol. 3; no. 1; pp. 67 - 75
• Main Authors: AOYAMA, Kazuhiro, KOGA, Tsuyoshi, IWATA, Yoshiharu, OKAMOTO, Kazuya, SATOH, Ryohei, MURATA, Hidenori, MORINAGA, Eiji
• Format: Journal Article
• Language: English; Japanese
• Published: Smart Processing Society for Materials, Environment & Energy (High Temperature Society of Japan) 20.01.2014
• Subjects: Automatic Optimization; DSM; SDSI-Cubic; System Design; Systems Engineering
• ISSN: 2186-702X; 2187-1337; 2187-1337
• DOI: 10.7791/jspmee.3.67

System is becoming large scale, complication, and diversification. Hence,an optimal product system design reflecting demands of society is required. To this problem, we built the system (named SDSI-Cubic). It creates optimization workflow order of the design process without any feedbacks based on the defined product system by SysML. However, the method of generating an optimization workflow automatically using DSM from SysML is not proposed. So, in this research, we propose the automatic generation method of the task DSM in consideration of optimization: new algorithm 1, by adding the dependency between design variables and objective functions to the conventional task DSM. Furthermore, we also propose the automatic method of generation an optimization workflow: new algorithm 2, by defining conversion rule from the partitioning result of this task DSM to an optimization workflow. And by performing automatic optimization of a cantilever using SDSI-Cubic incorporating these algorithms, we verified that it could optimize by using the algorithm to propose. As the future work, in order to solve a complicated problem efficiently, we need to consider the shortening method of optimization time by dividing an optimization problem.