Abstract: Aiming at low efficiency and unsatisfactory results while solving the large-scale problems, to take the wind farm layout problem with variable parameters as a study target, the strategy of hyper heuristics is designed and implemented. The paper selects the DE（Differential Evolution） algorithm and CMA-ES（Covariance Matrix Adaptation Evolution Strategy） algorithm as the low-level operators, and at the high level it uses the hyper heuristics algorithm to call the low-level operators to solve the wind farm layout problem under different complicated conditions. By contrast, the experiment data imply that the new strategy is more efficient and flexible.

Key words: hyper heuristic algorithm, wind farm layout, Differential Evolution（DE） algorithm, Covariance Matrix Adaptation Evolution Strategy（CMA-ES） algorithm

摘要： 针对大规模问题求解效率不高、结果不理想等问题，以影响参数多变的风力发电机布局问题为研究对象，设计并实现了超启发式算法策略，底层算子用差分进化（Differential Evolution，DE）算法和适应性协方差策略（Covariance Matrix Adaptation Evolution Strategy，CMA-ES）算法，高层策略用启发式调用策略选择底层算子求解在不同场景、不同风力参数等多种情况下的风力发电机布局情况。实验将权值选择策略与DE算法、CMA-ES算法和随机调度策略进行比较，最终数据表明该策略求解风力发电布局的效果远高于其他三种。

关键词: 超启发式算法, 风力发电机布局, 差分进化算法, 适应性协方差矩阵进化策略算法