To solve the problem that artificial bee colony algorithm is good at exploration and neglect exploitation, this paper proposes an improved artificial bee colony algorithm based on genetic model, which has been successfully applied to array synthesis. Firstly, the global optimal solution is introduced into the neighborhood search process to guide the bees to find the best nectar source thus to accelerate the convergence of the algorithm. Secondly, to avoid the local optimization of the algorithm, the exploitation ability of artificial bee colony algorithm must be improved. The evolutionary mechanism of genetic algorithm is used for reference, and a genetic model is established to carry out genetic operation on the honey source after adopting the optimal retention, to enrich the diversity of honey source. The improved artificial bee colony algorithm is tested on a set of widely used numerical functions, and the experimental data show that the proposed algorithm has strong competitiveness compared with other algorithms. Then, the algorithm is applied to the sparse optimization of the linear array to reduce the peak sidelobe level of the array. The optimization is compared with other algorithms under the same array constraints. The simulation results further prove the effectiveness of the algorithm.

The beam direction of the phased array antenna changes dynamically. Its electromagnetic emission characteristics show significant statistical laws, and the resources required for analysis and testing are numerous. In this paper, polynomial chaos expansion (PCE) is used to explore the statistical characteristics of the two-dimensional planar phased array antenna’s emission. The objective function of the phased array antenna is determined according to the pattern product theorem, and PCE is used to establish an equivalent surrogate model of the array’s emission characteristics. The paper starts from the phased array antenna composed of ideal point sources, and considers two typical cases that the main beam direction obeys the uniform distribution and the normal distribution. The statistical characteristics of the equivalent surrogate model are obtained by computer simulation. The results of the traditional Monte Carlo (MC) method are used as a reference to evaluate the effectiveness and reliability of the PCE method. At the end of the paper, the situation that the beam direction of the dipole phased array antenna obeys two typical distributions is briefly discussed. The comparison of simulation results shows that the PCE method can greatly reduce the sampling numbers of calculation while ensuring the accuracy of the results, and significantly improve the efficiency of analysis and testing.