Abstract: In order to realize the global optimization of the response target of the interdigital electrode sensor, a joint optimization method based on the approximate model of the improved goose optimization-least squares support vector machine(IGOOSE-LSSVM) and the non-dominated sorting whale optimization algorithm(NSWOA) is proposed. Firstly, the establishment principles of the IGOOSE-LSSVM approximation model and the NSWOA algorithm were analyzed, the finite element model of the interdigital electrode was constructed, and the structural parameters required for optimization were selected. Then, based on the electrode simulation model, the sample data set of the approximation model was constructed by Latin hypercubic sample sampling, and the sample data was inputted into the IGOOSE-LSSVM model for training to obtain the desired approximation model. Finally, the joint approximation model and the NSWOA algorithm are used to find the optimal parameter optimization combinations of the interdigital electrode with the optimization objectives of maximum output capacitance of the interdigital electrode and minimum size of the electrodes, and the results obtained are compared and verified. Experimental results demonstrate that the established approximation model effectively fits the electrode output capacitance and dimensional parameters, achieving correlation coefficients of 0.995 1 and 0.995 6 respectively. Compared to orthogonal experiments, the proposed hybrid optimization method achieves a 35.61% improvement in output capacitance, a 50.96% enhancement in capacitance response sensitivity, and a 43% reduction in electrode dimensions.

    Key words: interdigitated electrode; sensor; IGOOSE-LSSVM approximate model; non-dominated sorting whale optimization algorithm;parameter optimization