Definition of dp state: dp[s][p] means the maximum total value we can get from the p-th part by putting some amount of boxes starting from s.
And to calculate the number of distinct integers in a range I used a persistent segment tree. First I stored the previous occurance for each number. Then for any interval i to j I only need to see how many numbers in previous[i...j] are less than i. And that is where the persistent seg tree helped me.
So if I am not wrong then at the worst case the dp table will be fully filled so the complexity for the solve() function would be O(n*k*distinct_int_query) in this case it is O(n*k*log n). Building the segment tree at first will take O(n*log n). Afterwards updating for each element will take a total of O(n*log n)
So total complexity would be O(n*k*log n) right? Then why doesn't it pass? Why TLE???
P.S: Most likely I am wrong. So plz tell how to do the wrong thing right.