Research On Fast Antenna Design Based On Swarm Intelligence Optimization Algorithm

The property of antenna with electromagnetic radiation makes it able to act as a conversion device between electrical and electromagnetic signals,and its performance plays a crucial role in the quality of modern communication.With more and more complex antenna application scenarios,the demand for high performance antennas has become more and more urgent,and antenna design has become a very challenging task.The antenna design process has the disadvantages of high cost,time consuming and low efficiency based on electromagnetic simulation software,and the initial structure is not easy to determine,and the discrete antenna structure optimization is easy to fall into the "curse of dimensionality",etc.These problems are urgently needed to be solved.In this thesis,an extensive and in-depth research is carried out to improve the inefficiency of both continuous and discrete antenna design,and a fast and optimal antenna design strategy based on swarm intelligence optimization algorithm is proposed,the main work is as follows.1.In order to reduce the huge time cost brought by the expensive electromagnetic simulation software in the antenna optimization process,PSO-DBN antenna agent model is proposed in this thesis.For the problem that the number of deep belief networks(DBN)hidden layer nodes is not easily determined,the particle swarm optimization algorithm(PSO)is introduced to determine the initial structure of the network.Meanwhile,to enhance the sample richness,the Latin hypercube sampling(LHS)method is used to sample in the antenna parameter space,and finally the neural network is trained to construct the PSO-DBN antenna Surrogate model.Multi-objective design experiments of continuous-type parametric antenna structures are carried out using this model,and the design examples of small planar multi-band antenna verify the efficiency of the above antenna design scheme.2.In order to make the antenna structure have great design freedom and appear rich and diverse structural design solutions,this thesis uses antenna topology optimization design method.For the discrete space variables optimization prone to a series of problems caused by high dimensionality,the phase-amplitude modulation bat algorithm(P-AMBA)is proposed.The algorithm can map high-dimensional binary variables by optimizing only six-dimensional continuous variables,which not only can significantly reduce the optimization time,but also can improve the global search capability.Finally,the proposed algorithm is tested numerically for the 0/1 backpack problem,and a design example of a compact dual-band planar monopole antenna is given.The experimental results verify the superiority of P-AMBA for high-dimensional binary optimization problems and the feasibility of antenna topology optimization design.