Online deadline scheduling with preemption penalties

When ρ = 0, we give an optimal, O(Δ/log Δ)-competitive algorithm that, unlike previous algorithms, does not require knowledge of Δ. This settles an open problem mentioned in Ting (2008).