- Title
- Iterative train scheduling in networks with tree topologies: a case study for the Hunter Valley Coal Chain
- Creator
- Mendes, A.; Jackson, M.; Rocha de Paula, M.; Rojas, O.
- Relation
- 22nd International Congress on Modelling and Simulation (MODSIM2017). Proceedings of the 22nd International Congress on Modelling and Simulation (Hobart, Tas 03-08 December, 2017) p. 1337-1343
- Relation
- https://www.mssanz.org.au/modsim2017/
- Publisher
- The Modelling and Simulation Society of Australia and New Zealand (MSSANZ)
- Resource Type
- conference paper
- Date
- 2017
- Description
- This work describes an optimisation method based on genetic algorithms to generate train schedules for the rail network under the coordinating responsibility of the Hunter Valley Coal Chain Coordinator in NSW. The network connects 3 coal export terminals to 31 load points and haulage distances can extend up to 364 km. The scheduling problem consists of finding a high-quality schedule for trains travelling from a terminal to a load point and back, respecting all constraints imposed by the network itself and the operational environment. Those constraints refer to a mix of single and double tracks, limited parking facilities along the tracks, loading capacity at the load points, as well as minimum spacing (headway) between trains. The decision variables include the travel speeds at each section of the network and the amount of dwell time for each train at each parking facility along the route. To test our approach, a simplified model of the HVCCC network, with 3 terminals, 11 load points and 40 sections was used. The objective function is the minimization of the total travel times. A lower bound for that objective function was calculated with the trains travelling at maximum speed, and no constraints being applied. Three scenarios were tested, with 15, 30 and 60 trains; and with different configurations of the genetic algorithm. The results are presented in the form of a table with a number of statistics related to the solutions found, namely average travel time (with standard deviation), plus shortest and longest travel times, and CPU times. Relative to the lower bounds, the gaps for the average trip time range between 14% and 50%, depending of the problem size. These initial results are encouraging, considering the complexity of the system, the number and complexity of constraints, and the CPU time required by the method. Finally, in the discussion section we indicate possible paths of future research.
- Subject
- train scheduling; supply chain; coal transportation; optimisation; genetic algorithms
- Identifier
- http://hdl.handle.net/${Handle}
- Identifier
- uon:33846
- Identifier
- ISBN:9780987214379
- Language
- eng
- Reviewed
- Hits: 2075
- Visitors: 2050
- Downloads: 0
Thumbnail | File | Description | Size | Format |
---|