The need to schedule activities lies at the heart of many supply chain and logistics operations. Effective scheduling has long been critical to profitability. More recently, the impetus towards tighter delivery timeframes as an essential aspect of a logistics service has been growing strongly, as internet business drives increased consumer expectations of delivery services, and large online retailers see shorter delivery times as a serious competitive edge. But scheduling has long been a challenge for optimization. Discretization of time can yield strong integer programming models, but often, in practical settings, the large scale of these formulations prohibits their solution. Here we discuss recent work on two different types of logistics scheduling problem: (i) scheduling maintenance jobs, that shut down arcs in a network for the duration of the job, so as to maximize the total flow over time, and (ii) scheduling less-than-truckload truck dispatch operations so as to deliver goods within service time windows, while exploiting consolidation opportunities to minimize transport costs. We highlight the challenge of modeling time in these settings, and show, in the latter case, how properties of the problem permit much more efficient techniques for handling time than traditional approaches.