ROMJIST Volume 20, No. 3, 2017, pp. 241-255
Lucian N. VINTAN Towards Synergic Meta-Algorithmic Approaches in Complex Computing Systems
ABSTRACT: Today for almost any complex problem there are a set of state of the art refined powerful basic algorithms, quite effective in solving it. Frequently, there is not an ideal algorithm for a given complex research problem. Thus, for such complex applications it would be recommendable to use state of the art algorithms that will be concurrently orchestrated through a meta-algorithmic abstraction layer. In this paper we review this fertile idea. So, we analysed a meta-optimization method proposed in order to find the best (Pareto) individuals in a bi-objective space. Some qualitative and quantitative results were presented and explained. Also it was shown that meta-algorithms were efficient in documents' classi-fication domain. An effective adaptive meta-classification scheme for text documents was presented. Finally, it was shown that meta-algorithmic approaches are very effective in developing meta-predictors, too. A generic adaptive meta-prediction scheme was developed. In all these validation cases it was shown that the meta-algorithmic approaches involve useful synergism.KEYWORDS: Complex Computing Systems, Meta–Algorithm, Meta–Optimization, Multi- Objective (Pareto) Optimization, Meta–Classification, Meta–PredictionRead full text (pdf)