A case-based decision support model for the semiconductor assembly

Abstract

For a semiconductor package to have better performance, cost effectiveness, and better reliability, there are a variety of factors that should be considered such as packaging materials and process condition. To meet the minimum reliability level required by customers, trial and error tests are done and redone. In selection of the packaging materials for assembly, an experienced engineer's knowledge and memory of prior assembly experiences can be powerful aid. However, even for experts, it is not easy to make a decision. Case-based reasoning helps to resolve this issue: it provides support for decision makers when solving new problems on the basis of past experiences(i.e., previous cases). Case-based reasoning is a new technology that helps to solve problems by storing, retrieving and adapting past situations or cases. In case-based system a case is usually a specific problem that has been previously encountered and solved. Like most humans, when faced with a new problem, a case-based system compares the current problem with cases encountered in the past to determine if one of the earlier experiences can provide a solution. If a similar case exists, the solution to complex problems is found in matter of seconds. If the system retrieves a case that is similar but not entirely appropriate to the current problem, the case base system will provide a new solution from known cases. This new case will be added to the case base for future use. CBR works well in domains that are poorly understood because the case-based system does not need to know why a solution worked in the past. In this paper, preconditioning test results of semiconductor packages are predicted and proper die attach adhesive are selected by applying CBR, and the reults are very promising.