Application of IGWO Optimization Algorithm-based Interval Type-2 TSK FLS Method in the Soft Measurement of Chemical Process
Considering strong nonlinearity and complexity in the soft sensing modeling for chemical pro-cess,a IGWO-IT2 TSK FLS method was proposed based on interval Type-2 Takagi-Sugeno-Kang fuzzy logic system and combined with an improved grey wolf optimizer.Compared with the TSK fuzzy logic system,the IT2 TSK FLS method adeptly captured both intra-individual and inter-individual uncertainties.On the basis of the existing error back propagation(BP)algorithm training,the IGWO algorithm was further used to de-sign model's input and output so as to improve prediction performance of the model.Through improving the grey wolf optimization algorithm,the premature convergence judgment mechanism,nonlinear cosine adjust-ment strategy and Levy flight strategy were introduced to improve convergence speed of the algorithm and avoid its falling into local optimum.In addition,the IGWO-IT2 TSK FLS method was applied to the model-ing of the soft sensing examples for a debutane tower.Under the same conditions,having the IGWO-IT2 TSK FLS method compared with the type-1 TSK FLS method,BP algorithm,GA,differential evolution(DE),PSO,biogeography-based optimization(BBO)and grey wolf optimization(GWO)was implement-ed,respectively.The experimental results show that,the IGWO-IT2 TSK FLS method outperforms them in performance,including its effectiveness and application potential.
万方数据
维普
