Surface wave exploration has the advantages of high shallow resolution, small site limitations, and convenient construction, increasing its use in near-surface exploration. Dispersion curve inversion is an important step in surface wave exploration and is directly related to the acquisition of underground formation information. Similar to numerous geophysical inversion problems, dispersion curve inversion has multiparameter and multiextreme characteristics and is difficult to solve using a linear method. In this paper, a new dispersion curve inversion method based on the Archimedes optimization algorithm (AOA), namely the improved AOA (IAOA), is proposed. IAOA optimizes the population initialization based on AOA and adds the capability to automatically balance global search and local development performance in the iteration process, which enriches the population information in the later stage of AOA iteration. The algorithm is used to invert the noise and noise-free base-step scatter curves of the three theoretical models to test the performance of IAOA for dispersion curve inversion. In the inversion test of the theoretical model of noise-free data, the particle swarm optimization (PSO) algorithm and AOA were also tested in the same inversion test to compare the performances of the PSO, AOA, and IAOA algorithms. The results of the model test revealed that IAOA can stably and quickly converge to the optimal solution, and the inversion results have strong credibility, good noise immunity, and can be effectively used for dispersion curve inversion. Finally, the measured data from the Wyoming area of the United States were utilized to test the capability of IAOA to invert actual data. The inversion results indicated that IAOA has strong practicability and can obtain effective formation information.

面波勘探具有浅层分辨率高、场地限制小、施工方便等优点，越来越多地用于近地表勘探。频散曲线反演是面波勘探的重要步骤，直接关系到地下地层信息的获取。与众多地球物理反演问题类似，频散曲线反演具有多参数、多极值的特点，难以用线性方法求解。本文提出了一种新的基于阿基米德优化算法（AOA）的频散曲线反演方法，即改进的AOA（IAOA）。IAOA在AOA的基础上优化了种群初始化，在迭代过程中增加了自动平衡全局搜索和局部开发性能的能力，丰富了AOA迭代后期的种群信息。该算法用于反演三种理论模型的噪声和无噪声基步散布曲线，以测试IAOA对散布曲线反演的性能。在无噪声数据理论模型的反演测试中，粒子群优化算法（PSO）和AOA也在同一反演测试中进行了测试，比较了PSO、AOA和IAOA算法的性能。模型试验结果表明，IAOA能够稳定快速收敛到最优解，反演结果可信度强，抗噪性好，可有效用于频散曲线反演。最后，利用美国怀俄明地区的实测数据，测试了IAOA反演实际数据的能力。反演结果表明，IAOA具有较强的实用性，能够获取有效的地层信息。