最近做了SCOI2014GDOI2014

觉得智商还是有问题

看到一种牛逼的决策单调优化

def work(L, R, dl, dr) :
    m = (L + R) / 2
    for i = dl to dr :
        update g[m] using i
    if L == R : return 
    let dm be the optimal decision of m
    work(L, m - 1, dl, dm)
    work(m + 1, R, dm, dr)