Compared to the previous approach of, our algorithm improves the complexity from an exponential dependence on T to a nearly quadratic dependence, under the condition R < 1. And let’s take a peek ...