Recently, many kinds of computer aided engineering (CAE) codes have become available. However, selecting a suitable code for effective and efficient computation is not easy for users with no knowledge of CAE or no interest in the features of computer systems. Additionally, it is quite difficult to determine appropriate computational parameters even if a well-established CAE code is used. In this study, we propose a framework for selecting parameters appropriate to distributed resources such as computers, databases and knowledge. Two prototype systems are implemented on a PC cluster and a Grid. Numerical experiments show that our proposed framework will be able to recommend appropriate computational parameter to users, and to extract unknown rules in the computations.
展开▼