A Systematic Test Generation Approach to Achieving VariousTest Qualities for LSIs Based on Mathematical Programming

这项工作的重点是数学规划（特别是整数规划）作为一个测试生成框架，可以灵活地生成所需的大规模集成电路（LSI）的各种类型的测试。已经设计了几种方法来制定不同的测试生成问题的整数规划模型和启发式地解决它们。这个框架允许我们用一个强大的现成的整数规划求解器来解决这些问题，而不需要开发任何专门的新技术。研究结果可以在低成本的情况下提高LSI的可靠性。

This work focused attention on mathematical programming (especially, integer programming) as a test generation framework that can flexibly generate various types of tests required for large-scale integrated circuits (LSIs). Several methods have been devised to formulate integer programming models for different test generation problems and to heuristically solve them. This framework allows us to solve these problems with a powerful off-the-shelf solver for integer programming, without developing any newtechniques dedicated to them. The results can lead to enhancing the reliability of LSIs at low cost.