Abstract: Membrane systems are models of computation which are inspired by some basic features of biological membranes. Transition P systems are very simple models. Many hardware and software architectures have been proposed for implementing them. In particular, there are implementations in cluster of processors, in microcontrollers and in specialized hardware. This work proposes an analysis of the P system in order to be able to reduce the execution time of a given evolution step. We present a solution for improving the time of working out the active rules subset of a membrane. This task is critical for the entire evolution process efficiency because it is performed inside each membrane in every evolution step. Therefore, we propose to carry out a static analysis over the P system. The collected information is used for obtaining a decision tree for each membrane. During the execution time of the P system, active rules of a membrane will be determined as a result of a classification problem from the corresponding decision tree. By incorporating decision trees for this task, we will notice some improvements.

Keywords: Decision Tree, ID3, Active Rules, Transition P System

ACM Classification Keywords: I.2.6 Learning – Decision Tree; D.1.m Miscellaneous – Natural Computing

Link:

IMPROVING ACTIVE RULES PERFORMANCE IN NEW P SYSTEM COMMUNICATION ARCHITECTURES

Juan Alberto de Frutos, Luis Fernández, Carmen Luengo, Alberto Arteta

http://www.foibg.com/ijitk/ijitk-vol04/ijitk04-1-p01.pdf