Solution 1 DP

• t-complexity: $O(n)$

• s-complexity: $O(C)$ C = 40001

n is in range [-10000, 10000], n-diff is in range [-20000, 20000]. and move to positive as index

• state definition: dp[n] is longest subsequence ended with n

• state transfer: dp[n] = dp[n-diff] + 1

• init value: None

• return value: max(dp[arr[i]])