Logistics

SMECS – AI-based ETA Forecasting in Intermodal Transportation Networks

Development of predictive models for proactive disruption management in intermodal logistics networks using artificial intelligence

Website with online demonstrator: https://smecs-eta.de/

Project content

Motivation

Due to the increasing dynamics and complexity of the procurement and sales markets, the susceptibility to disruptions in maritime transport networks is growing. This results in longer response times to disruptive events, while at the same time the demands on delivery service and delivery time are constantly increasing. Intelligent monitoring and control systems, with which automatic identification is possible even before disruptions occur and proactive measures for the players involved are indicated in the sense of a prescriptive analytics approach, represent tools with which these challenges can be efficiently met.

 

Goals

The research project SMECS (Smart Event Forecast for Seaports) contributes to the realization of agile transport networks by developing a model that enables a proactive detection of cross-actor disturbances and thus a targeted risk management. The focus is on investigations of process events affecting the ETA (Estimated Time of Arrival) of the means of transport used within multimodal and international container transports for export in the port and port hinterland.

The aim of the project is to improve the forecasting of delays and terminal congestion by integrating all the players involved and to provide suitable alternative actions for transport control in the event of an incident, thus increasing the efficiency and robustness of planning and service provision for the overall system.

 

Solution approach

Based on a system analysis of the maritime transport chain, process events with disruptive character as well as their causes and consequences are determined and transferred into an impact chain model. Both operational and external factors such as weather and traffic are taken into account.

By identifying relevant predictors in the corresponding data sources and applying suitable procedures, a forecast model for real-time ETA calculation in export traffic is developed. The focus is initially on train transports and will subsequently be extended to include other freight transport modes. By aggregating the determined information, volume flows are also forecasted so that statements on terminal utilization as well as long-term developments are possible.

By subsequently linking the potential disruptions with identified measures for avoiding disruptions or reducing the consequential effects, a risk management system is developed that identifies actor-specific action alternatives depending on the process events.

Results, dates etc.

Project successfully completed in February 2020

On February 13, 2020, a final conference for the SMECS project ("Smart Event Forecast for Seaports") took place at the Kühne Logistics University in Hamburg. At this event, the results of 2.5 years of project work were presented to representatives from business, research and politics.

Under the leadership of the Department of Logistics at TU Berlin, an intelligent assistance system for the maritime transport chain was developed in the SMCES project by applying artificial intelligence methods (machine learning). The realized IT system predicts with a high accuracy the time of arrival (ETA) of intermodal container transports from the shipper of goods to the seaport and thereby proactively detects conflicts in the adherence to the planned process flow. By providing disruption-related measures, the actors in the logistics chain are enabled to carry out the transports reliably and efficiently.

For the ETA forecast, various data from more than 15 IT systems were integrated on historical transport processes and a variety of operational and environmental factors that have an influence on the individual transport and handling processes. This included data on the infrastructure, resources and disruptions of the logistics companies involved, as well as external data sources such as weather and traffic.

The project was funded as part of the "Innovative Port Technologies" (IHATEC) initiative of the German Federal Ministry of Transport and Digital Infrastructure (BMVI). On the TU Berlin side, the project was handled by Peter Poschmann and Manuel Weinke. In addition to the project partners KLU and DB Cargo AG, SMECS found great support from other companies by providing operational information and data.

 

Project results

The (interim) results of the project were published via various media and are available here, among others:

Project Organization

Project sponsor:

Project executing agency:

  • TÜV Rheinland Consulting GmbH

Collaboration partner:

  • Technische Universität Berlin (TU Berlin), Fachgebiet Logistik (Konsortialführer)
  • Kühne Logistics University – Wissenschaftliche Hochschule für Logistik und Unternehmensführung (KLU)
  • DB Cargo AG

Associated partner:

  • boxXpress.de GmbH
  • Dakosy Datenkommunikationssystem AG
  • DB Netz AG 
  • Hamburg Süd - Hamburg Südamerikanische Dampfschifffahrts-Gesellschaft (HSDG)
  • Kühne + Nagel (AG & Co.) KG
  • Lübecker Hafen-Gesellschaft mbH
  • Metrans a.s.
  • TFG Transfracht Internat. Gesellschaft für kombinierten Güterverkehr mb
  • Verein Hamburger Spediteure e.V. 
  • et al. 

Project duration: 

01.09.2017 - 29.02.2020

Contact person knowledge management

M.Sc.

Jonas Brands

Research Associate

brands@logistik.tu-berlin.de

+49 30 314-28438

Organization name Logistics
Office H 90
Building Hauptgebäude (H)
Room H 9103
Office hoursby appointment