静的学習における効率的な間接含意の発見と保存について [in Japanese] On Efficient Identification and Preservation of Indirect Implications in Static Learning [in Japanese]
Search this Article
Author(s)
Abstract
経路活性化法に基づくATPG(自動テストパターン生成)では,静的学習で見つけた間接含意を含意操作に利用することが処理の効率化に役立つ.大規模回路に対してはATPGの効率化の要求はなおも強くある一方で,得られる間接含意が多すぎて,保存に必要となる記憶領域が増大することが問題になることがある.本論文では,これまでの静的学習では得られなかった間接含意を見つけることのできる新しい学習規則を提案する.また,冗長な間接含意の保存を避けるような静的学習の処理方法について考察する.
In automatic test pattern generation based on the path sensitization method, it is useful for an implication procedure to utilize indirect implications derived by static learning. While it still has been required to speed-up ATPG for large circuits, there is a problem that there are too many indirect implications to store. In this paper we propose a new learning criterion to find indirect implications that cannot be found by the previous method. Also we consider a method of static learning to avoid preserving redundant indirect implications.
Journal
-
- IEICE technical report. Dependable computing
-
IEICE technical report. Dependable computing 102(479), 25-28, 2002-11-21
The Institute of Electronics, Information and Communication Engineers