|The Spring 1998 INFORMS National Meeting was held in Montreal, April 26-29, 1998. The cluster chair was Mike Beeby at 514-399-8924, fax 514-399-5579, e-mail at Beeby@cn.ca.|
|Presentations by Line Capacity Model vendors were made. This panel consited of:|
When we reconvened, a panel of railroad users had an
opportunity to make brief presentations or simply ask questions of the
This panel consited of:
|RASIG founder Ajith Wijeratne presented outgoing chair Mike Gorman with a plaque thanking him for his distinguished service.|
Abstract: In their bid to become more competitive, cost efficient and customer responsive many North American railways are moving towards the "Operating to Plan" or a scheduled railway philosophy. Canadian National's efforts in this direction were described. This includes the OTP architecture and components (forecasting, train planning, resource validation), and a discussion of our experience with these system components.
Abstract: As part of the CN's "Operating to Plan" initiative, much emphasis was placed on the forecasting of expected train workload. This workload includes traffic originating on CN's lines and traffic received at interchanges from other railroads. The trials, failures, and triumphs of the forecasting layman were presented.
Abstract: We described a fleet (locomotives and cars) assignment system for passenger transportation, called RAIL-WAYS. It can be used to study what-if scenarios, to compute the equipment cycles, and to update these cycles in order to respond to short-term car demand changes. RAIL-WAYS is based on a column generation approach. Computational results were presented.
We proposed a formulation and an exact algorithm, based on Benders decomposition, for an equipment assignment problem arising in the context of railway passenger transportation. Adaptations to different scenarios were discussed, together with implementation considerations. Computational experiments carried out on real-life instances indicated that the algorithm is highly efficient.
Abstract: We presented a dynamic programming approximation for dynamic management of locomotives. Our method is able to handle a variety of complex operating strategies and the characteristics of major locomotive types. The method explicitly handles uncertainties in forecasted tonnages, and can optimize over extended horizons with fast response times.
Abstract: The locomotive scheduling system provides at minimum cost sufficient motive power to pull all the trains scheduled by the railway company while satisfying the available locomotive fleet power and the locomotive maintenance planning. The numerical results are provided for an acyclic problem over a one week horizon, using data obtained from the company CN North America. This problem involves 2000 trains and 1300 locomotives.
Abstract: Summary of a practitioners review of capacity and service reliability on a 750 mile corridor, including effects of changing priority, P/W ratios, plant enhancement, track maintenance, line stoppages, through historic review, and desecrate event simulation.RCM is used to analyze the traffic capacity of rail line segments. Train delay statistics and time-distance plots permit evaluation of a given plant configuration and signalling system. RCM can model single or double track and has fairly sophisticated meet pass logic which governs which train will wait where, in the case of a conflict.
Abstract: Parametric approach to identifying line capacity by focusing on key
plant, traffic & operating factors that drive train delay.
Yard Capacity Model
Canadian National Railway.
Abstract: Assessment of Yard capacity with respect to traffic volumes, in & out bound process and throughput, and an assessment of the significance of each.