Parallel Computing in SAS®: Genetic Algorithms Application
Alejandro Correa Bahnsen, Darwin Amezquita, Andres M. Gonzalez, Banco Colpatria · 2012
Genetic Algorithms is a very powerful optimization technique that can be used in a wide variety of problems. But unfortunately the performance of this methodology relies heavily on computer power. We used Genetic Algorithms to select the architecture of a Multi-Layer Perceptron Neural Network and even though results indicated that it improves the predictive power of credit risk models, it is also very time consuming and computationally expensive. Because of this, we implemented a version of Parallel Genetic Algorithms in SAS® using PROC CONNECT procedure. Results show that parallel computing can drastically reduce the total execution time of the genetic algorithm.