Transportation Cost Assessment by Means of a Monte Carlo Simulation in a Transshipment Model

• Main Authors: Gordana Dukić, Darko Dukić, Ivana Alerić
• Format: Article
• Language: English
• Published: University of Zagreb, Faculty of Transport and Traffic Sciences 2008-09-01
• Series: Promet (Zagreb)
• Subjects: transportation costs; stochastic transshipment model; Monte Carlo simulation; quantitative methods; confidence interval; !-distribution; computer analysis
• Online Access: http://www.fpz.unizg.hr/traffic/index.php/PROMTT/article/view/1014
• ISSN: 0353-5320; 1848-4069

The task of transport management is to organize the transportof goods from a number of sources to a number of destinationswith minimum total costs. The basic transportation modelassumes direct transport of goods from a source to a destinationwith constant unit transportation costs. In practice, however,goods are frequently transported through several transientpoints where they need to be transshipped. In such circumstancestransport planning and organization become increasinglycomplex. This is especially noticeable in water transport.Most of the issues are directly connected to port operations, asthey are the transshipment hubs. Since transportation is under anumber of influences, in today 's turbulent operating conditionsthe assumption on fixed unit transportation costs cannot betaken as realistic. In order to improve decision making in thetransportation domain, this paper will present a stochastictransshipment model in which cost estimate is based on MonteCarlo simulation. Simulated values of unit costs are used to devisean adequate linear programming model, the solving ofwhich determines the values of total minimum transportationcosts. After repeating the simulation for a sufficient number oftimes, the distribution of total minimum costs can be formed,which is the basis for the pertinent confidence interval estimation.It follows that the design, testing and application of thepresented model requires a combination of quantitative optimizationmethods, simulation and elements of inferential statistics,all with the support of computer and adequate software.