Computation of weighted sums of rewards for concurrent MDPs