Identification of Thermal Process Based on Relevance Vector Machine and Firefly Algorithm

Abstract

Thermal process usually has the characteristics of nonlinearity and uncertainty, so it is difficult to establish the nonlinear model by the traditional method of describing the dynamic mathematical model of thermal process and accurately implement optimal control of thermal process. This paper provides an identification method of thermal process based on relevance vector machine and firefly algorithm, and the modeling results are compared with those by minimal resource allocation network and support vector regression. The simulation results show that the method is effective, and the model has high accuracy and can be directly used in control algorithm based on the model.