算法收敛性分析：相对误差曲线图解读

Revision: Algorithm convergence refers to whether an algorithm can approximate the optimal solution or achieve the desired accuracy within a finite number of iterations. When an algorithm converges, the computed results gradually approach the true values, and the relative error decreases. Therefore, the relative error curve can visually demonstrate the process and speed of algorithm convergence. If the algorithm converges, the error should gradually decrease and stabilize as the number of iterations increases. Figure 1 on page 6 of the paper shows the relative error change curves for three sub-problems. All three relative error curves gradually decrease and eventually converge to a small value, indicating algorithm convergence.