Methodology 方法论

1. Problem Formulation 问题建模

We consider a multilayer network G = (V,E).
我们考虑一个多层网络 G = (V,E)。

Edge capacity:

P(C_e = k) = p_k

Reliability:

R = P(max_flow ≥ D)

2. Monte Carlo Estimation 蒙特卡洛估计

R ≈ (1/N) Σ I(flow_i ≥ D)

3. Attention-Based Learning 基于注意力的学习

The model learns mapping:

f: Edge Features → Reliability

模型学习从边特征到可靠性的映射。