Skip to main content
SpringerLink
Log in

A disaggregate stochastic freight transport model for Sweden

  • Published:
Transportation Aims and scope Submit manuscript

Abstract

This paper presents estimation results for models of transport chain and shipment size choice, as well as an implementation of the estimated disaggregate models (for two commodity groups), in the context of the national freight transport model for Sweden. The new model is a disaggregate and stochastic (logit) model, whereas the existing Swedish national model is deterministic. One advantage of the new approach is that it bases the underlying behavior of shippers on a stronger empirical foundation (that is micro-data from the Swedish Commodity Flow Survey, CFS). Another advantage is that it overcomes a well-known disadvantage of deterministic models that lead to implausibly large responses to changes in scenario or policy variables. Although estimation and implementation of aggregate stochastic models were done before, in the context of a national freight transport forecasting model, we think this is the first implementation of disaggregate freight transport chain and shipment size models estimated on choice data for individual shipments, certainly in Europe. We carried out a number of model runs with both versions of the implemented model to compare elasticities and found that transport cost and time elasticities for tonne-km are smaller (in absolute values) in the disaggregate stochastic model than in their deterministic counterparts.

This is a preview of subscription content, log in via an institution to check access.

Access this article

Log in via an institution

Price excludes VAT (USA)
Tax calculation will be finalised during checkout.

Instant access to the full article PDF.

Institutional subscriptions

Fig. 1
Fig. 2

Source: Ben-Akiva and de Jong (2013)

Fig. 3
Fig. 4
Fig. 5

Similar content being viewed by others

Notes

  1. See Chow et al. (2010) for a comprehensive review of freight forecast models elsewhere.

  2. A transport chain is defined here as a series of modes that are all used to transport a shipment from the sender to the receiver (e.g. road–sea–road).

  3. “SAMGODS review” section gives a brief overview of the national freight transport model for Sweden, SAMGODS. At the core of this model is the ADA model framework first suggested by de Jong and Ben-Akiva (2007). It starts with an aggregate model for the determination of flows of goods between production (P) zones and consumption (C) zones. After this comes a disaggregate “logistics” model, that based on PC flows produces OD (origin–destination) flows for the network assignment which is the third phase (aggregate again). For example, A PC flow that uses the transport chain road–sea–road between the production and consumption locations contributes to three OD flows (one for each of the modes in the chain).

  4. Moreover, models for shipment size and mode choice have been developed based on the French ECHO dataset at the shipment level (Combes 2010).

  5. A partial exception is that the Danish national freight model contains a module for the choice of mode to cross the Fehmarn Belt screenline that uses a random utility model estimated on disaggregate data (including stated preference SP surveys in the Fehmarn Belt corridor). Other transport chains, however, for example in Denmark, are handled by a deterministic logistics model (Ben-Akiva and de Jong 2013, section 4.6).

  6. We are not comparing different network assignment techniques in this paper (both methods rely on the same skims from unimodal networks which yield input variables for the allocation to transport chain and shipment size that is being studied here).

  7. See http://www.trafikverket.se/contentassets/23a269d514d24920ad445881d724811f/filer/vfu_2004_2005.pdf for details.

  8. In this study, as in most previous studies, we consider the weight of shipment size as an endogenous variable. However, we note that shipment volume (in m3) is also an important factor, which shippers consider jointly with mode choice decisions. We cannot model shipment volume because our data set, the Swedish CFS, does not contain this information.

  9. We note that there could be correlations between alternatives, especially given that there are alternatives that have a transport chain (or a shipment size) in common. More complicated nesting structures can be tried in mixed logit and multivariate probit models, but these model types have very long run times, especially on large data sets as we have here.

  10. In the CFS a shipment is defined as a unique delivery of goods with the same commodity code to/from the local unit or to/from a particular recipient/supplier (SIKA 2004).

  11. We defined transport chain alternatives based on their frequency in the CFS. Transport chains that occurred with a frequency of 96 or higher were considered as possible choice options.

  12. The dependent (choice) variable (Ui) in Eq. 1 is defined based on the classification on Table 1.

  13. The maximum gross weight of the trucks is 60 tonnes in Sweden and Finland compared to 40 tonnes in most other European countries.

  14. Akin to de Jong and Ben-Akiva (2007) a recent study by Zhao et al (2015) developed a freight temporal assignment model where disaggregate methods are used to assign aggregate annual flows to aggregate daily flows. We note there are other approaches to simulating freight flows at the national or broad regional levels using different cost functions, micro-simulation and agent-based approaches or direct-demand modeling in various countries, which are reviewed by Chow et al. (2010) (especially US studies) and de Jong et al. (2013) and Liedtke (2009) (especially European studies). Wisetjindawat et al. (2007) also developed a micro-based freight model for the Tokyo Metropolitan area.

  15. Tonne-km in Sweden is the sum of the domestic transports and the domestic parts of international transports that are carried out in Sweden.

  16. Furthermore, there can also be changes in shipment size in both models as a result of cost changes.

References

  • Abate, M., de Jong, G.C.: The optimal shipment size and truck size choice-the allocation of trucks across hauls. Transp. Res. Part A 59(1), 262–277 (2014)

    Google Scholar 

  • Abate, M., Vierth, I., de Jong, G.: Joint econometric models of freight transport chain and shipment size choice. In: CTS Working Paper 2014, Stockholm, p. 9 (2014)

  • Abdelwahab, W., Sargious, M.: Modelling the demand for freight transport: a new approach. J. Transp. Econ. Policy 26(1), 49–70 (1992). http://www.jtep.com

    Google Scholar 

  • Ben-Akiva, M., de Jong, G.C.: The aggregate–disaggregate–aggregate (ADA) freight model system. In: Ben-Akiva, M.E., Meersman, H., van de Voorde, E. (eds.) Freight Transport Modelling. Emerald, Bingley (2013)

    Chapter  Google Scholar 

  • Bovenkerk, M.: SMILE+: the new and improved Dutch national freight model system. In: Paper Presented at the European Transport Conference, Strasbourg (2005)

  • Caspersen, E., Johansen, B.G., Hovi, I.B., de Jong, G.C.: https://www.toi.no/publikasjoner/nasjonal-godstransportmodell-fra-et-deterministisk-rammeverk-til-en-stokastisk-modell-article34088-8.html (2016)

  • Chow, J.Y., Yang, C.H., Regan, A.C.: State-of-the art of freight forecast modeling: lessons learned and the road ahead. Transportation 37(6), 1011–1030 (2010)

    Article  Google Scholar 

  • Combes, P.F.: Estimation of the economic order quantity model using the ECHO shipment data base. In: Paper Presented at the European Transport Conference, Glasgow (2010)

  • Combes, F., Tavasszy, L.A.: Inventory theory, mode choice and network structure in freight transport. Eur. J. Transp. Infrastruct. Res. 16(1), 2016 (2016)

    Google Scholar 

  • de Jong, G.C., Ben-Akiva, M.: A micro-simulation model of shipment size. Transp. Res. B 41(9), 950–965 (2007)

    Article  Google Scholar 

  • de Jong, G., Schroten, A., van Essen, H., Otten, M., Bucci, P.: Price Sensitivity of European Road Freight Transport—Towards a Better Understanding of Existing Results—a Report for Transport and Environment. Significance & Delft, Delft (2010)

    Google Scholar 

  • de Jong, G.C., Burgess, A., Tavasszy, L., Versteegh, R., de Bok, M., Schmorak, N.: Distribution and modal split models for freight transport in The Netherlands. In: Paper Presented at ETC 2011, Glasgow (2011)

  • de Jong, G.C., Vierth, I., Tavasszy, L., Ben-Akiva, M.E.: Recent developments in national and international freight transport models within Europe. Transportation 40–2, 347–371 (2013)

    Article  Google Scholar 

  • Grebe, S., de Jong, G.C., de Bok, M., van Houwe, P., Borremans, D.: The strategic Flemish freight model at the intersection of policy issues and available data. In: Paper Presented at ETC 2016, Barcelona (2016)

  • Holguín-Veras, J.: Revealed preference analysis of commercial vehicle choice process. J. Transp. Eng. 128(4), 336–346 (2002)

    Article  Google Scholar 

  • Inaba, F.S., Wallace, N.E.: Spatial price competition and the demand for freight transportation. Rev. Econ. Stat. 71(4), 614–625 (1989)

    Article  Google Scholar 

  • Jensen, A.F., Thorhauge, M., de Jong, G.C., Rich, J., Dekker, T., Johnson, D., Ojeda Cabral, M., Bates, J., Anker Nielsen, O.: A model for freight transport chain choice in Europe. In: Paper Presented at the HEART Conference, Delft, (2016)

  • Johnson, D., de Jong, G.C.: Shippers’ response to transport cost and time and model specification in freight mode and shipment size choice. In: Proceedings of the 2nd International Choice Modeling Conference ICMC 2011, University of Leeds, United Kingdom, 4–6 July (2011)

  • Jourguin, B., Tavasszy, L., Duan, L.: On the generalized cost—demand elasticity of intermodal container transport. Eur. J. Transp. Infrastruct. Res. 14–4, 362–374 (2014)

    Google Scholar 

  • Liedtke, G.: Principles of micro-behavior commodity transport modeling. Transp. Res. Part E 45, 795–809 (2009)

    Article  Google Scholar 

  • Lloret-Batlle, R., Combes, F.: Estimation of an inventory theoretical model of mode choice in freight transport. Transp. Res. Rec. J. Transp. Res. Board 2378, 13–21 (2013)

    Article  Google Scholar 

  • McFadden, D., Winston, C., Boersch-Supan, A.: Joint estimation of freight transportation decisions under non-random sampling. In: Daugherty, A. (ed.) Analytical Studies in Transport Economics, pp. 137–157. Cambridge University Press, Cambridge (1985)

    Google Scholar 

  • Rich, J., Holmblad, P., Hansen, C.: A weighted logit freight mode-choice model. Transp. Res. Part E 45, 1006–1019 (2009)

    Article  Google Scholar 

  • RSG: User guide and model documentation: agent-based supply chain modeling tool. Chicago Metropolitan Agency for Planning. Resource Systems Group, Chicago (2015)

    Google Scholar 

  • Significance: Method Report—Logistics Model in the Swedish National Freight Model System, Deliverable for Trafikverket, Report D6B, Project 15017. Significance, The Hague (2015)

    Google Scholar 

  • SIKA: The Swedish National Freight Model: a critical review and an outline of the way ahead, Samplan 2004:1. SIKA, Stockholm (2004)

    Google Scholar 

  • Tavasszy, L.A., Smeenk, B., Ruijgrok, C.J.: A DSS for modelling logistics chains in freight transport systems analysis. Int. Trans. Oper. Res. 5(6), 447–459 (1998)

    Article  Google Scholar 

  • Vierth, I., Jonsson, L., Karlsson, R., Abate, M.: Konkurrensytta land—sjö för svenska godstransporter. VTI (VTI Rapport 822/2014) (2014)

  • Windisch, E., de Jong, G.C, van Nes, R.: A disaggregate freight transport model of transport chain choice and shipment size choice. In: Paper Presented at ETC 2010, Glasgow (2010)

  • Wisetjindawat, W., Sano, K., Matsumoto, S., Raothanachonkun, P.: Micro-simulation model for modeling freight agents interactions in urban freight movements. In: Paper Presented at 86th Annual Meeting of the Transportation Research Board, Washington DC (2007)

  • Zhao, M., Chow, J.Y., Ritchie, S.G.: An inventory-based simulation model for annual-to-daily temporal freight assignment. Transp. Res. Part E Logist. Transp. Rev. 79, 83–101 (2015)

    Article  Google Scholar 

Download references

Author information

Authors and Affiliations

  1. The World Bank, 1818 H St. NW, Washington, DC, 20433, USA

    Megersa Abate

  2. Swedish National Road and Transport Research Institute, Linköping, Sweden

    Inge Vierth & Rune Karlsson

  3. Institute for Transport Studies, University of Leeds, Leeds, UK

    Gerard de Jong

  4. Significance, Den Haag, The Netherlands

    Gerard de Jong & Jaap Baak

Authors
  1. Megersa Abate
    View author publications

    You can also search for this author in PubMed Google Scholar

  2. Inge Vierth
    View author publications

    You can also search for this author in PubMed Google Scholar

  3. Rune Karlsson
    View author publications

    You can also search for this author in PubMed Google Scholar

  4. Gerard de Jong
    View author publications

    You can also search for this author in PubMed Google Scholar

  5. Jaap Baak
    View author publications

    You can also search for this author in PubMed Google Scholar

Corresponding author

Correspondence to Megersa Abate.

Rights and permissions

Reprints and permissions

About this article

Check for updates. Verify currency and authenticity via CrossMark

Cite this article

Abate, M., Vierth, I., Karlsson, R. et al. A disaggregate stochastic freight transport model for Sweden. Transportation 46, 671–696 (2019). https://doi.org/10.1007/s11116-018-9856-9

Download citation

  • Published:

  • Issue Date:

  • DOI: https://doi.org/10.1007/s11116-018-9856-9

Keywords

Search

Navigation