Training datasets of size 1K, 3K, and 9K for four utility function variants

The datasets include 1K, 3K, and 9K training data points that are used for training utility-change prediction models in systems with ground truth from four different mathematical complexities: Linear utility function Saturating utility function Discontinuous utility function Combined utility function ( combining the above three variants)本数据集包含1K、3K与9K三种规模的训练数据点，用于训练各类系统中的效用变化预测模型，该类模型的真实标签（ground truth）源自四种不同数学复杂度的效用函数：线性效用函数（Linear utility function）、饱和效用函数（Saturating utility function）、非连续效用函数（Discontinuous utility function）以及组合效用函数（Combined utility function，整合上述三类效用函数变体）。