It is well established that in many scenarios there is no single solver that will provide optimal performance across a wide range of problem instances. Taking ad- vantage of this observation, research into algorithm se- lection is designed to help identify the best approach for each problem at hand. This segregation is usually based on carefully constructed features, designed to quickly present the overall structure of the instance as a constant size numeric vector. Based on these features, a plethora of machine learning techniques can be utilized to pre- dict the appropriate solver to execute, leading to sig- nificant improvements over relying solely on any one solver. However, being manually constructed, the cre- ation of good features is an arduous task requiring a great deal of knowledge of the problem domain of in- terest. To alleviate this costly yet crucial step, this paper presents an automated methodology for producing an informative set of features utilizing a deep neural net- work. We show that the presented approach completely automates the algorithm selection pipeline and is able to achieve significantly better performance than a sin- gle best solver across multiple problem domains.},
...
...
@@ -138,3 +134,9 @@
publisher="IEEE Press",
year={2009}
}
@article{GNNandSat,
author={Benedikt Bünz and
Matthew Lamm},
title={Graph Neural Network and Boolean Satisfiability},
