Faktorförsökbaserad diskret händelsestyrd simulering av trafikflöden under långsiktiga prognoser: En fallstudie vid Göteborgs Hamn
Independent thesis Advanced level (professional degree), 20 credits / 30 HE creditsStudent thesis
World tradeis steadily increasing and the most common way of transporting goods is oversea. In the European Union, 90% of all external trade is through the use of maritimetrade. As the trade volume increase, so does the volume of the vessels. Thismeans that a greater amount of goods arrive simultaneously when a ship moors ata harbour. In order to handle this a harbour has to increase their efficiencycontinuously. To be able to accept the largest vessels a harbour has to beahead of the evolution of seafaring vessels, this requires forecasts with longtime horizons.The problemwith forecasts is, the longer the forecast the greater the uncertainty. Whenthe uncertainty is too high, scenarios are often developed instead of usingmathematical models. A tool shown to be highly effective at testing scenariosis simulation. Furthermore simulation is a cost effective way to investigatethe outcome of a potential investment.Standardtrial and error methods of investigating effects of a simulation is often timeconsuming and inefficient. By implementing design of experiments (DoE) to thesimulation process the naivety of trial and error can be avoided. The carmanufacturing industry as well as the health-care sector has already integratedsimulation with DoE to great success. In the generally capital intensiveinfrastructure projects, DoE and simulation has not seen the same penetration. Thisbrings us to the purpose of this thesis, namely; How the combination of simulation and design of experiments can beexpressed in a infrastructure project with a long time horizon.This thesiswas conducted at the Port of Gothenburg and simulated how different queue timesand length was affected by different scenarios and factors. In order to handlethese factors and scenarios, which amounted to 160, DoE and simulation wasapplied.Thequantitative data used was acquired through the use of archived material concerningpassages through the ports in May and February 2015. Participant observationswere also used. The qualitative data used was attained through interviews, bothstructured and focused, as well as participant observations. The qualitativedata was used to create an understanding of how the real system was expressed. Thisin order to articulate it in a simulation model, with the software Simul8. Quantitative data was used tofit distributions and other data for the simulation.Data fromthe simulation was analysed with software DesignExpert. With it the dependent variables, queue length and queue time, wereidentified. The factor that had the greatest effect on the dependent variableswas the capacity inside the terminal. The implementation of arrival timewindows was shown to have a statistically significant effect on the dependentvariables. Therecommendations to the Port of Gothenburg are as follows. Seeing as theadministration of the terminals is not their responsibility and thus is outsideof their jurisdiction, the recommendation is to implement an arrival time window system. This willensure a smoother flow of traffic to terminals. If the arrival time windowsystem is introduced, an easy way to increase the capacity is to increase theopening hours of the terminals. However, in the future, the capacity of the terminalmust be increased.Instead ofimplementing a hard cap on the number of vehicles, a fee can be introduced.This fee should vary during the day; charging more during the most activehours. This would dissuade drivers to arrive during the busiest hours.Furthermore, Port Entry would need 59 parking spaces in order to guarantee every one a place, 37 in 75% of cases. An introduction of arrival time windows would decrease the cost by 45 million Swedish crowns annually for the carters by reducing their waiting time.
Place, publisher, year, edition, pages
2015. , 83 p.
Teknik, Simulering, diskret händelsestyrd simulering, flöde, trafikflöde, prognos, långsiktiga prognoser, Göteborgs Hamn, faktorförsök, simulation, Discret event simulation, Forecast, Design of Experiments
IdentifiersURN: urn:nbn:se:ltu:diva-53260Local ID: a4a87a55-6048-4638-9336-7b7a585946b6OAI: oai:DiVA.org:ltu-53260DiVA: diva2:1026634
Subject / course
Student thesis, at least 30 credits
Industrial and Management Engineering, master's level
Validerat; 20150611 (global_studentproject_submitter)2016-10-042016-10-04Bibliographically approved