Abstract
Line planning is one of the important steps in the public transit planning process. It is a difficult combinatorial problem with an enormous search space. For networks from practice, which are typically large and have more complex constraints, it takes an unreasonable amount of time to find good solutions with the current methods. When the objective is to minimize passenger travel time, most of the calculation time is spent on solving the passenger routing sub-problem. This paper proposes a local evaluation technique to significantly speed up this most critical component. A change is assumed to have its biggest impact on passengers already travelling close to the change. This idea is implemented by putting a grid over the network and only re-evaluating the part of the network around the change, while still taking into account all passengers. This results in a smaller computation time for each passenger routing, with only a limited loss of quality, and it allows more iterations. On a number of benchmark instances, the results obtained by the grid approach are compared to the same algorithm without any local evaluation. The grid approach is significantly faster and allows to perform 5–20 times more evaluations per second, while having a minimal loss of quality per evaluation. When the available calculation time is limited, the grid based approach obtains the best solutions. On the larger benchmark instances, the grid based approach can use more complex neighborhoods resulting in even better solutions. The method is very flexible and can be integrated in many line planning algorithms and probably even in algorithms for other network problems.
Similar content being viewed by others
Data availability
References
Ahmed L, Mumford C, Kheiri A (2019) Solving urban transit route design problem using selection hyper-heuristics. Eur J Oper Res 274(2):545–559. https://doi.org/10.1016/j.ejor.2018.10.022
Baaj MH, Mahmassani HS (1991) An AI-based approach for transit route system planning and design. J Adv Transp 25(2):187–209. https://doi.org/10.1002/atr.5670250205
Baaj MH, Mahmassani HS (1995) Hybrid route generation heuristic algorithm for the design of transit networks. Transp Res Part C Emerg Technol 3(1):31–50. https://doi.org/10.1016/0968-090X(94)00011-S
Borndörfer R, Grötschel M, Pfetsch ME (2007) A column-generation approach to line planning in public transport. Transp Sci 41(1):123–132
Bull S, Larsen J, Lusby RM, Rezanova NJ (2019) Optimising the travel time of a line plan. 4OR 17(3):225–259. https://doi.org/10.1007/s10288-018-0391-5
Canca D, De-Los-Santos A, Laporte G, Mesa JA (2019) Integrated railway rapid transit network design and line planning problem with maximum profit. Transp Res Part E Logist Transp Rev 127:1–30. https://doi.org/10.1016/j.tre.2019.04.007
Ceder A (2016) Public transit planning and operation: modeling, practice and behaviour, 2nd edn. CRC Press
Ceder A, Wilson NHM (1986) Bus network design. Transp Res Part B Methodol 20(4):331–344. https://doi.org/10.1016/0191-2615(86)90047-0
Chakroborty P, Wivedi T (2002) Optimal route network design for transit systems using genetic algorithms. Eng Optim 34(1):83–100. https://doi.org/10.1080/03052150210909
Constantin I, Florian M (1995) Optimizing frequencies in a transit network: a nonlinear bi-level programming approach. Int Trans Oper Res 2(2):149–164. https://doi.org/10.1016/0969-6016(94)00023-M
Dijkstra EW (1959) A note on two problems in connexion with graphs. Numer Math 1(1):269–271. https://doi.org/10.1007/BF01386390
Durán-Micco J, Vermeir E, Vansteenwegen P (2019) Considering emissions in the transit network design and frequency setting problem with a heterogeneous fleet. Eur J Oper Res 282(2):580–592. https://doi.org/10.1016/j.ejor.2019.09.050
Fan L, Mumford CL (2010) A metaheuristic approach to the urban transit routing problem. J Heuristics 16(3):353–372. https://doi.org/10.1007/s10732-008-9089-8
Farahani RZ, Miandoabchi E, Szeto WY, Rashidi H (2013) A review of urban transportation network design problems. Eur J Oper Res 229(2):281–302. https://doi.org/10.1016/j.ejor.2013.01.001
Fu H, Nie L, Meng L, Sperry BR, He Z (2015) A hierarchical line planning approach for a large-scale high speed rail network: The China case. Transp Res Part A Policy Practice 75:61–83. https://doi.org/10.1016/j.tra.2015.03.013
Gattermann P, Schiewe A, Schöbel A (2016) An iterative approach for integrated planning in public transportation. In: 9th triennial symposium on transportation analysis
Gentile G, Noekel K (eds) (2016) Modelling public transport passenger flows in the era of intelligent transport systems: COST action TU1004 (TransITS). Springer, Berlin. https://doi.org/10.1007/978-3-319-25082-3
Goerigk M, Schmidt M (2017) Line planning with user-optimal route choice. Eur J Oper Res 259(2):424–436. https://doi.org/10.1016/j.ejor.2016.10.034
Guan JF, Yang H, Wirasinghe SC (2006) Simultaneous optimization of transit line configuration and passenger line assignment. Transp Res Part B Methodol 40(10):885–902. https://doi.org/10.1016/j.trb.2005.12.003
Guihaire V, Hao J-K (2008) Transit network design and scheduling: a global review. Transp Res Part A Policy Practice 42(10):1251–1273. https://doi.org/10.1016/j.tra.2008.03.011
Ibarra-Rojas OJ, Delgado F, Giesen R, Muñoz JC (2015) Planning, operation, and control of bus transport systems: a literature review. Transp Res Part B Methodol 77:38–75. https://doi.org/10.1016/j.trb.2015.03.002
Iliopoulou C, Kepaptsoglou K, Vlahogianni E (2019) Metaheuristics for the transit route network design problem: a review and comparative analysis. Public Transport 11:487–521. https://doi.org/10.1007/s12469-019-00211-2
Islam KA, Moosa IM, Mobin J, Nayeem MA, Rahman MS (2019) A heuristic aided stochastic beam search algorithm for solving the transit network design problem. Swarm Evol Comput 46:154–170. https://doi.org/10.1016/j.swevo.2019.02.007
John MP, Mumford CL, Lewis R (2014) An improved multi-objective algorithm for the urban transit routing problem. In: Blum C, Ochoa G (eds) Evolutionary computation in combinatorial optimization, vol 8600, Springer, Berlin, pp. 49–60. https://doi.org/10.1007/978-3-662-44320-0_5
Johnson DB (1977) Efficient algorithms for shortest paths in sparse networks. J ACM 24(1):1–13. https://doi.org/10.1145/321992.321993
Kepaptsoglou K, Karlaftis M (2009) Transit route network design problem: review. J Transp Eng 135(8):491–505. https://doi.org/10.1061/(ASCE)0733-947X(2009)135:8(491)
Kılıç F, Gök M (2014) A demand based route generation algorithm for public transit network design. Comput Oper Res 51:21–29. https://doi.org/10.1016/j.cor.2014.05.001
Laporte G, Ortega FA, Pozo MA, Puerto J (2017) Multi-objective integration of timetables, vehicle schedules and user routings in a transit network. Transp Res Part B Methodol 98:94–112. https://doi.org/10.1016/j.trb.2016.12.018
Liu Y, Bunker J, Ferreira L (2010) Transit users’ route-choice modelling in transit assignment: a review. Transp Rev 30(6):753–769. https://doi.org/10.1080/01441641003744261
Lusby RM, Larsen J, Bull S (2018) A survey on robustness in railway planning. Eur J Oper Res 266(1):1–15. https://doi.org/10.1016/j.ejor.2017.07.044
Mandl CE (1979) Applied network optimization. Academic Press
Mandl CE (1980) Evaluation and optimization of urban public transportation networks. Eur J Oper Res 5(6):396–404. https://doi.org/10.1016/0377-2217(80)90126-5
Mauttone A, Urquhart ME (2009) A multi-objective metaheuristic approach for the transit network design problem. Public Transport 1(4):253–273. https://doi.org/10.1007/s12469-010-0016-7
Mumford CL (2013a) The data, results and supplementary material can be downloaded from: http://users.cs.cf.ac.uk/C.L.Mumford/Research%20Topics/UTRP/CEC2013Supp/
Mumford CL (2013b) New heuristic and evolutionary operators for the multi-objective urban transit routing problem. IEEE Congress Evolut Comput 2013:939–946. https://doi.org/10.1109/CEC.2013.6557668
Nayeem MA, Rahman MdK, Rahman MS (2014) Transit network design by genetic algorithm with elitism. Transp Res Part C Emerg Technol 46:30–45. https://doi.org/10.1016/j.trc.2014.05.002
Nikolić M, Teodorović D (2013) Transit network design by bee colony optimization. Expert Syst Appl 40(15):5945–5955. https://doi.org/10.1016/j.eswa.2013.05.002
Pattnaik SB, Mohan S, Tom VM (1998) Urban bus transit route network design using genetic algorithm. J Transp Eng 124(4):368–375. https://doi.org/10.1061/(ASCE)0733-947X(1998)124:4(368)
Patz A (1925) Die richtige Auswahl von Verkehrslinien bei grossen Strassenbahnnetzen. Verkehrstechnik 50:51
Schmidt ME (2014) Integrating routing decisions in public transportation problems. Springer, New York. https://doi.org/10.1007/978-1-4614-9566-6
Schmidt M, Schöbel A (2015) The complexity of integrating passenger routing decisions in public transportation models. Networks 65(3):228–243. https://doi.org/10.1002/net.21600
Schöbel A (2012) Line planning in public transportation: models and methods. Or Spectrum 34(3):491–510. https://doi.org/10.1007/s00291-011-0251-6
Schöbel A, Scholl S (2006) Line planning with minimal traveling time. In: 5th Workshop on algorithmic methods and models for optimization of railways (ATMOS’05), vol 16. https://doi.org/10.4230/OASIcs.ATMOS.2005.660
Sörensen K (2015) Metaheuristics-the metaphor exposed. Int Trans Oper Res 22(1):3–18. https://doi.org/10.1111/itor.12001
Wardrop J (1952) Some theoretical aspects of road traffic research. In: Proceedings of the institute of civil engineers, part II, vol 1, pp 325–378. https://doi.org/10.1680/ipeds.1952.11362
Williams RR (2018) Faster all-pairs shortest paths via circuit complexity. SIAM J Comput 47(5):1965–1985. https://doi.org/10.1137/15M1024524
Zhao F, Zeng X (2006) Simulated annealing-genetic algorithm for transit network optimization. J Comput Civ Eng 20(1):57–68. https://doi.org/10.1061/(ASCE)0887-3801
Acknowledgements
This research is funded by the Research Foundation Flanders project G.0853.16N.
Author information
Authors and Affiliations
Ethics declarations
Conflict of interest
The authors declare that they have no conflict of interest.
Ethical standard
We comply with all ethical responsibilities of an author.
Additional information
Publisher's Note
Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations.
Appendix
Appendix
1.1 Partial origin–destination matrices
In this appendix, we explain in more detail how the calculation of the partial OD matrices is implemented. While the concept itself is quite simple, we noticed the implementation can be quite tricky and complicated.
To calculate the total travel time in a cut of the network, both the extended transit network and the OD matrix are necessary. Each cut of 3 by 3 cells has its own OD matrix, called a partial OD matrix. The calculation of the partial OD matrices is responsible for most of the overhead in the grid method. For efficiency purposes, these matrices are calculated at the same time when evaluating a newly implemented solution. The only thing needed to calculate these partial OD matrices is the original OD matrix, the routes the passengers will take with the current line plan and the network information.
Each passenger is assumed to enter and leave each cut at a fixed location, based on their route in the current solution. Only their path in between these two points can change, based on the local search moves considered. Starting or ending their trip in a node present in the cut counts as, respectively, entering or leaving the network. In Fig. 3, a passenger travelling from node 0 to node 9 will enter the highlighted cut (3, 2) in node 2b. As mentioned before, 2b is a border node representing the entrance of a bus line. From a demand perspective, this means that any demand from 0 (or 1 or 2) to 9 gets converted into demand from 2b to 9. All passengers that enter the cut while on this bus line will enter the cut in node 2b and thus will create demand starting (or ending) in 2b. It should be noted that a single passenger route might need to be split up in multiple routes. This means it is possible that a shortest path of a passenger enters (and leaves) a cut twice.
Calculating all these partial OD matrices might seem computationally expensive, but actually all partial OD matrices can be calculated at the same time. Each cut and partial OD matrix can be identified by the coordinates of its central cell. The cut in the bottom left corner of a network is then identified as cut (1, 1) and the cut next to it is cut (2, 1). Because of the overlap between the 3 by 3 cells, each node is present in multiple cuts.
When constructing all partial OD matrices, each shortest path (in the complete network) is considered one by one and only once. This is done by moving through the shortest path and determining in which cuts each node along the shortest path is located. Whenever a new cut is entered, the previous node is recorded as the entry node. Remember from Sect. 2.2 that during the construction of the extended network, the first node outside the cut is included as a border point. That is the reason why the previous node is recorded and not the current. When a cut is left, the same process is executed. Now the current node is recorded as the exit node. This process is then applied to all OD pairs in the network. This results in a list of entry and exit nodes for each cut and every OD pair. The partial OD matrix of a cut can now be constructed by adding the demand from the complete OD matrix to each entry and exit pair in the partial OD matrix. This means the partial OD matrices contain both physical bus nodes (from the original network) and line nodes (entry and exit nodes).
This might seem a complicated process, but the result is that every OD-pair of which the shortest path in the current solution is going through a certain cut, is taken into account when a move is evaluated in this cut, by ‘moving’ this OD-pair to an artificial entry and exit node at the border of the cut. This approach ensures that a local evaluation serves as an upper bound to the travel time. The real impact of a change (on the entire network) can only be better than the impact estimated by the local evaluation method in a cut. This is because the routing is completely optimized within the cut of the network. The only possible improvements, not considered in the cut, are when a change makes a new route possible, previously not travelling through the cut, or when the impact of a change is less negative because an alternative route (outside the cut) can compensate part of the disadvantage estimated. The actual routes calculated by the local evaluation are possible, hence they are a possible solution of a full evaluation and thus the result of the full evaluation cannot be worse.
In the current implementation, every partial OD matrix contains an entry for each node in the original extended network. This requires a lot of memory. Since these are all very sparse matrices its size can be heavily reduced. Only the nodes that have any demand, have to be present in the partial OD matrix. Note that a list with node ids that links this limited OD matrix to the original is probably required. Just doing this will massively limit the memory use. By cleverly constructing the extended network of a cut it might be possible to avoid having to link the partial OD nodes to the original. Since we have not run into any memory problems we did not implement this, but we mention it because it is likely necessary when computing larger instances.
Rights and permissions
About this article
Cite this article
Vermeir, E., Durán-Micco, J. & Vansteenwegen, P. The grid based approach, a fast local evaluation technique for line planning. 4OR-Q J Oper Res 20, 603–635 (2022). https://doi.org/10.1007/s10288-021-00490-1
Received:
Revised:
Accepted:
Published:
Issue Date:
DOI: https://doi.org/10.1007/s10288-021-00490-1