The globally convergent convexification numerical method is constructed for a Coefficient Inverse Problem for the Mean Field Games System. A coefficient characterizing the global interaction term is recovered from the single measurement data. In particular, a new Carleman estimate for the Volterra integral operator is proven, and it stronger than the previously known one. Numerical results demonstrate accurate reconstructions from noisy data.

翻译：针对平均场博弈系统中的系数反问题，本文构建了全局收敛的凸化数值方法。通过单次测量数据反演表征全局相互作用项的系数。特别地，证明了Volterra积分算子的新型卡尔曼估计，其强度优于已知结果。数值实验表明，该方法能够从含噪数据中实现精确重构。