Boolean Matching with Abstraction and Dynamic Learning
碩士 === 臺灣大學 === 電子工程學研究所 === 98 === Boolean matching determines whether two given Boolean functions can be identical to each other under permutation and/or negation of their input and output variables. In this thesis, we focus on the computation kernel of Boolean matching and propose a complete gene...
| Main Authors: | , |
|---|---|
| Other Authors: | |
| Format: | Others |
| Language: | en_US |
| Published: |
2010
|
| Online Access: | http://ndltd.ncl.edu.tw/handle/63864700278674748385 |
| Summary: | 碩士 === 臺灣大學 === 電子工程學研究所 === 98 === Boolean matching determines whether two given Boolean functions can be identical to each other under permutation and/or negation of their input and output variables. In this thesis, we focus on the computation kernel of Boolean matching and propose a complete generic framework. We formulate the Boolean matching problem as SAT solving, and effectively prune infeasible matching solutions through conflict-driven learning and abstraction. Partial assignment reduction is applied to strengthen the power of learning. Our approach is capable of being easily integrated with signature-based techniques and applies them as preprocessing for quick search space reduction. Our framework is applicable for general Boolean matching problem, even for incompletely specified functions. The experimental results show the generality and scalability of our framework.
|
|---|