基于量子粒子群算法的自适应随机共振方法研究

摘要

In order to enhance the usefulness of the theory of stochastic resonance in the areas of weak signal detection, a new method based on quantum particle swarm optimization is proposed to conquer with the problem of adaptive stochastic resonance. First, the problem of adaptive stochastic resonance is converted into the problem of multi-parameter optimization. Then simulation experiments are conducted respectively under a Langevin system and Duffing oscillator system. At the same time, Point detection method is chosen as the comparative test in the Langevin system. While in the Duffing system, the optimization results are compared with those from the Langevin system directly. Results show that the method based on quantum particle swarm optimization is obviously superior to the point detection method and optimization result in the Duffing oscillator is better than that from Langevin system under the same condition. Besides, it is also found that the lower the SNR of input signal, the more effective the quantum particle swarm optimization is. Finally, the regularity of optimization results of the stochastic resonance system parameters is summarized.%为提升随机共振理论在微弱信号检测领域中的实用性，以随机共振系统参数为研究对象，提出了基于量子粒子群算法的自适应随机共振方法。首先将自适应随机共振问题转化为多参数并行寻优问题，然后分别在Langevin系统和Duffing振子系统下进行仿真实验。在Langevin系统中，将量子粒子群算法和描点法进行了寻优结果对比；在Duffing振子系统中， Duffing振子系统的寻优结果则直接与Langevin系统的寻优结果进行了对比。实验结果表明：在寻优结果和寻优效率上，基于量子粒子群算法的自适应随机共振方法要明显高于描点法；在相同条件下， Duffing振子系统的寻优结果要优于Langevin系统的寻优结果；在两种系统下，输入信号信噪比越低就越能体现出量子粒子群算法的优越性。最后还对随机共振系统参数的寻优结果进行了规律性总结。