为解决复杂地形下风电场微观选址优化难度大、耗时长的问题，提出一种基于代理模型辅助的复杂地形风电场微观选址多目标优化方法。首先，考虑复杂地形起伏大的地理特征，计算崎岖度指数，将地面平整度进行数值量化，并对崎岖度过大的点进行约束处理；其次，建立三维多风向下尾流叠加计算发电量的数学模型，构建三维地形集电线路拓扑优化代理模型，检验代理模型的预测精度，表明可替代集电线路拓扑优化的大量计算，有效提高计算效率，然后采用多目标离散状态转移算法，获得复杂地形微观选址多目标优化设计的Pareto前沿；最后，以中国新疆某实际复杂地形风电场为例，实现复杂地形风电场多目标微观选址，并将该算法与单目标优化结果进行比较。仿真结果表明，在考虑复杂地形因素特点的情况下，代理模型辅助的多目标离散状态转移算法能在优化年发电量的前提下减少总电缆铺设长度，降低工程建设投资，为工程实际提供更多布局方案。

In order to solve the problem of difficult and time-consuming micro-site optimization of wind farms under complex terrain, a multi-objective optimization method for micro-site selection of wind farms with complex terrain assisted by proxy model is proposed. Firstly, considering the geographical features of complex terrain with large undulations, the ruggedness index is calculated, the ground flatness is numerically quantified, and the points with excessive ruggedness are constrained. Secondly, a mathematical model of three-dimensional windy downward wake superposition calculation of power generation is established, a three-dimensional terrain collector line topology optimization agent model is constructed, and the prediction accuracy of the proxy model is verified, which shows that it can replace a large number of calculations for collector line topology optimization and effectively improve the computing efficiency. Finally, taking a real complex terrain wind farm in Xinjiang, China as an example, multi-objective micro-site selection of complex terrain wind farm is realized, and the algorithm is compared with the single-objective optimization results. The simulation results show that the multi-objective discrete state transfer algorithm assisted by the proxy model can reduce the total cable laying length, reduce the investment in engineering construction, and provide more layout schemes for the actual project under the premise of optimizing the annual power generation.