A hybrid Taguchi binary particle swarm optimization (HTBPSO) is proposed in this letter for electromagnetic optimization problems. The proposed algorithm combines the better properties of Taguchi method for updating the optimum positions of particles and improving the local search ability. To avoid premature convergence during the solution process, cat-fish effect operation is performed to alter the last 10% worse particle positions. Three examples, including the designs of linear and planar thinned arrays and planar tri-band mono-pole, are presented to demonstrate the efficiency of HTBPSO. As expected, experimental results show that HTBPSO achieves not only better solutions but also a faster convergence speed compared with the other two algorithms.

The sine cosine algorithm (SCA) is a newly emerging optimization algorithm. It is easy for sine cosine algorithm (SCA) to sink into premature of the algorithm and obtain a slower convergence rate when solving the complicated optimization problems, especially highly ill-posed problems. A novel modified sine cosine algorithm (MSCA) is put forward for solving the optimization problems. To our limited knowledge, the linear searching path and empirical parameter have not been applied to the related improved sine cosine algorithm. The proposed MSCA improves the search path of original SCA by introducing linear searching path and empirical parameter, effectively avoiding sinking into the local optimal. In addition, the proposed algorithm changes the definition of convergence factor. Two kinds of tests, including 23 benchmark functions test and actual engineering problem tests are adopted to prove the performance of the MSCA. In addition, the performance of the proposed MSCA is compared with SCA by using benchmark functions on different dimensional (D = 30, 50,100 and 500). As expected, the result of comparisons show that the proposed MSCA can better avoid the local optima than both the SCA and other population-based algorithms. And MSCA can obtain the faster convergence than SCA on different dimensions.

The spoof surface plasmon polaritons (SSPPs) structure can be used as a sensor in THz region for the biosensing. The accuracy of resonance and amplitude for sensor is very important for biosensing. The momentum matching of SSPPs determines the resonance position and the gap distance determines the amplitude. For the biomolecular sensing, the sample is positioned between the prism base and the SSPPs structure. The momentum matching condition at the current study does not consider the effect of sample refractive index and the resonance position has a significant error. Here the correction is made to the momentum matching condition which considers the effect of the sample refractive index. A comparative study of surface plasmon resonance (SPR) sensing performance based on frequency and angle variations shows that the sensing sensitivity for frequency region is superior to that of angle region; in the meanwhile, as an application of biosensors, we have detected different types of brain lesions in the frequency range. Furthermore, the reflection amplitude is related to gap size between the prism and SSPPs. The relationship of gap size and reflection amplitude is studied. By using the relationship between gap size and reflection amplitude, the amplitudes at different frequencies or incident angles for different refractivities have the same reflection dips compared to the other published results. The simulation is performed and the results proved the theory.

Highly sensitive detection of enhanced terahertz (THz) fingerprint absorption spectrum of trace-amount tiny molecules is essential for biosensing. THz surface plasmon resonance (SPR) sensors based on Otto prism-coupled attenuated total reflection (OPC-ATR) configuration have been recognized as a promising technology in biomedical detection applications. However, THz-SPR sensors based on the traditional OPC-ATR configuration have long been associated with low sensitivity, poor tunability, low refractive index resolution, large sample consumption, and lack of fingerprint analysis. Here, we propose an enhanced tunable high-sensitivity and trace-amount THz-SPR biosensor based on a composite periodic groove structure (CPGS). The elaborate geometric design of the spoof surface plasmon polaritons (SSPPs) metasurface increases the number of electromagnetic hot spots on the surface of the CPGS, improves the near-field enhancement effect of SSPPs, and enhances the interaction between THz wave and the sample. The results show that the sensitivity (S), figure of merit (FOM) and Q-factor (Q) can be increased to 6.55 THz/RIU, 4234.06 1/RIU and 629.28, respectively, when the refractive index range of the sample to measure is between 1 and 1.05 with the resolution $1.54\times {10}^{-5}$ RIU. Moreover, by making use of the high structural tunability of CPGS, the best sensitivity (SPR frequency shift) can be obtained when the resonant frequency of the metamaterial approaches the biological molecule oscillation. These advantages make CPGS a strong candidate for the high-sensitivity detection of trace-amount biochemical samples.

To avoid being easily trapped in a local optimal solution of Taguchi's method, an improved random Taguchi's method (IRTM) based on swarm intelligence and dynamic reduced rate (DRR) function is presented in this letter. The DRR is linked to the design parameter's percentage contribution toward performance variance obtained from the analysis of variance. To demonstrate the efficiency of the IRTM, four test function benchmarks and the design of an aperiodic planar array are investigated. Moreover, the results of the IRTM are compared with that of the classic meta-heuristic algorithms. From the comparison results, the IRTM is more efficient in that better performance could be obtained with less experiment times.

High-resolution microscopy technique is of significant importance for studying nanomaterials. It is necessary to understand the near-field interaction between the probe and substrate materials in order to get the fine structure of the nanomaterial in the subwavelength scale. The numerical methods such as FDTD, FEM, and MoM are inefficient for the SNOM problems because of the illness of the impedance matrix. The analytic method can only be used for some simple objects such as sphere. Here, a quasianalytical method is developed, in which the analytic formula is refined to adapt to various shapes of the probe approaching the curve of SNOM. By this way, it is helpful in comparing the performance of different probes and giving us a direction to design a new type probe in SNOM. As an application, the developed method is used to study the contrast in the SNOM for the interface between the two different surfaces that have different materials and topography.