Santa Fe Railway's intermodal business unit has developed an innovative approach to designing its service offerings. The operating-plan model (OPM) minimizes the schedule-related costs of service subject to rail-operating capabilities while meeting customers' expectations for service. The algorithm produces a weekly train timetable and assigns traffic to trains. Because of the problem's size and complexity, a combination of genetic and tabu searches is used to search for successively better operating plans. The OPM shows the potential to improve global service by four percent while reducing costs by six percent over the existing operating plan. Santa Fe Intermodal has realized savings by applying the OPM to more narrowly focused problems.