Sort the rewards array first.

If we decide to apply some rewards, it's always optimal to apply them in order.

The transition is given by: <code>dp[i][j] = dp[i - 1][j − rewardValues[i]]</code> if <code>j − rewardValues[i] < rewardValues[i]</code>.

Note that the dp array is a boolean array. We just need 1 bit per element, so we can use a bitset or something similar. We just need a "stream" of bits and apply bitwise operations to optimize the computations by a constant factor.