By Topic

Heterogeneous Catalysis Modeling Strategy Integrating Support Vector Regression and Chaotic Particle Swarm Optimization Algorithm

Sign In

Cookies must be enabled to login.After enabling cookies , please use refresh or reload or ctrl+f5 on the browser for the login options.

Formats Non-Member Member
$31 $13
Learn how you can qualify for the best price for this item!
Become an IEEE Member or Subscribe to
IEEE Xplore for exclusive pricing!
close button

puzzle piece

IEEE membership options for an individual and IEEE Xplore subscriptions for an organization offer the most affordable access to essential journal articles, conference papers, standards, eBooks, and eLearning courses.

Learn more about:

IEEE membership

IEEE Xplore subscriptions

4 Author(s)
Xiaoxia Han ; Coll. of Inf. Eng., Taiyuan Univ. of Technol., Taiyuan, China ; Gang Xie ; Keming Xie ; Jun Ren

A new heterogeneous catalysis modeling methodology, namely support vector regression (SVR) and chaotic particle swarm optimization algorithm (CPSO) was presented, for catalyst compositional models and catalytic reaction mechanism models, for reducing both high temporal costs and financial costs, and accelerating the process of industrialization synthesis of dimethyl ether (DME). In the SVR-CPSO approach, on the basis of characteristics of non-linearity uncertainty, higher-dimension, chaos and small quantity of training data, a support vector regression model was constructed for correlating process data comprising values of input variables of catalyst compositional, operating conditions and output variables of performance of catalyst (catalyst selectivity and conversion). In order to improve predictive accuracy and generalization ability of SVR models, CPSO was adopted to automatically determine the optimal hyper-parameters for SVR, which presented to avoid the premature phenomenon and the shortcoming of local optimum in PSO method, owns high effective and high precision and easy to use compared with GA. As a new strategy for heterogeneous catalysis modeling, the major advantage of SVR-CPSO strategy is that modeling can be conducted exclusively from the historic small sample space data wherein the detailed knowledge of process phenomenology (reaction mechanism, rate constants, etc.) is not required and difficult to get. Another advantage is that it has avoided the blindness and contingency of the traditional catalyst ? trial and error? method. Finally, this algorithm was verified experimentally to be feasible. The predicted performance numbers of catalyst of the SVR-CPSO model has been compared with the experimental values of the SVR-GA model. The result shows that the SVR-CPSO is an available data mining techniques; provides better prediction capability and is therefore considered as a promising alternative method for forecasting nonlinear dynamic system under the c- ircumstances of small sample.

Published in:

Computational Intelligence and Security, 2009. CIS '09. International Conference on  (Volume:2 )

Date of Conference:

11-14 Dec. 2009