Optimization of fleet mix - Case study for Wallenius Wilhelmsen Logistics Trond Johnsen 02 Oct 2012 Norsk Marinteknisk Forskningsinstitutt
Main planning levels in RoRo shipping Strategic Planning Fleet Size and Mix Tactical Planning Fleet Deployment Operational Planning Stowage Planning
Project MARFLIX Maritime fleet size and mix optimisation Strategic planning of the size and mix of the fleet of vessels Types of ships to include in the fleet, their sizes and the number of ships of each size A key issue in ship-owner / ship operator planning Known generically in OR as Fleet Size and Mix Problems (FSMP)?? MARFLIX Objective: Develop and test methods (and models) for improved decision support for fleet size and mix decisions in a maritime transport system context; to improve the use of state-of-research (OR theory and applications) in applied maritime FSMP decision making.
MARFLIX project info Budget 16 MNOK over 2010 2013 75 % financing from NFR Academia PM: Prof Asbjørnslett at NTNU IMT R&D Post Doc: Jørgen Rakke (IMT) 2 PhD candidates: Giovanni Pantuso (NTNU IØT) and Rikard Bakkehaug (NTNU IMT) 6-8 Master thesis WWL and DNV as industry partners Development and testing of methods and tools for solving fleet size and mix problems Research performed by MARINTEK, with participation from DNV pronavis
Case study - WWL
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Decision variables A suggested fleet with a given number of vessels within each vessel type that satisfy all constraints and represent the lowest objetive function (min cost) A deployment of the fleet, explaining how many trips will be made on each trade by a certain vessel type with a certain speed
Model objective Objective function minimize cost! Fuel costs pr vessel type At ports and at sea HFO MGO Port costs pr vessel type (and trade) Panama Suez Average port cost Time charter costs
Constraints Fleet constraints Fleet fixation (min, max) Flexibility of vessel capacity Vessel capacity requirements between subsets of regions Trades from A to B with a given Time duration depending on speed (and number of ports) Frequency requirement pr trade Vessel capacity requirement pr trade Between regions Between subsets of regions Vessel types with Product capacities pr vessel Fuel consumption pr vessel (and speed) Compatibility characteristics between vessels and trades
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Mathematical model
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Excel user interface
Model implemented in Mosel, solved with Xpress solver
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Model output Text files
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Cargo segments
The optimization process Real Life Decision Problem Identification, limitations and simplifications Simplified problem formulation Mathematical modeling Optimization model/ MP model Assessment Analysis Analysis and Verification Result Optimization method Solution process Decision and implementation
Decision support and implementation Solves large problems in less than 10 seconds: 60 vessels of 20 different types 25 trade lines 5 year planning horizon Added value for the shipping company: Decision support for the strategic fleet renewal process Improved utilisation of fleet capacity with optimized deployment and speed selection Supports comprehensive scenario based analysis
Scenario New trade AF->US
Scenario Future ECA areas
Remaining challenges MARFLIX started late 2010, and the case activity will conclude early 2013 Remaining challenges: Handling of uncertainty in the future cargo volumes in order to give robust fleet size and mix advises Validation of the expected performance of the fleet suggested by the model Balancing the trade-off between accuracy in the data input and user friendliness of the model
Thank you for your attention! Questions? Trond Johnsen Research Manager, Logistics & Operations Research MARINTEK Dep. of Maritime Transport Systems trond.johnsen@marintek.sintef.no Mobile: +47 92 21 64 89 Norsk Marinteknisk Forskningsinstitutt