The following is an English translation of the Norwegian project description, which was revised during spring 2013 after notice from the Research Council of Norway that funding had been granted.
Background and Motivation
An important challenge within the offshore industry and sea traffic preparedness is how to plan and optimise the positioning of a fleet of vessels in a dynamic, everchanging environment such that operating costs and risks related to health, the environment, and safety are reduced, whilst services and profit are improved. In Norway, the Norwegian Coastal Administration (NCA) is
responsible for the tug vessel preparedness along the coast of Norway. By monitoring and assessing the current and future (predicted) state of factors such as the traffic situation, ocean currents, wave heights, and weather conditions, the NCA attempts to plan for and position their fleet of tugs in a best possible manner. Currently there exists no method, algorithm, or computer programme that can do this objectively and fact-based. Instead, the decision-making and positioning of the tug fleet is done ad hoc on a subjective basis by human operators at a vessel traffic service (VTS) centre in the town of Vardø.
Basing such important decisions solely on human knowledge and experience is prone to error, especially with the projected increase in ship traffic and the complex situations that may arise and that can sometimes be difficult to read and fully understand. Thus objective and calculating methods implemented in a computer programme can be a useful decision-support tool for the VTS operators. Moreover, such methods may also be transferrable to, and be of aid for, fleet optimisation in the offshore industry. With the planned growth and expansion of oil and gas fields in the High North, the ship traffic will see a huge increase during the coming years, for example with an increasing number of platform supply vessels (PSVs).
The development of novel methods for tug fleet optimisation (TFO) in the DRAMA project has the potential of leading to the development of software prototypes and implementation of actual decision-support tools with real-world application. The socioeconomic benefits are potentially great in terms of a reduced number of accidents that can lead to loss of human lives and environmental damage. In addition, there are benefits in terms of faster and safer marine operations and less transportation expenses.
Project Group Members
The project group consists of the following core people employed at the Faculty of Engineering and Natural Sciences (AIR) and Faculty of Technology and Maritime Operations (AMO) at Aalesund University College (AAUC):
- Associate Professor Robin T. Bye (project head), AIR-ICT
- Professor Hans Georg Schaathun, AIR-ICT
- Associate Professor Siebe van Albada, AIR-Natural Sciences
- Senior engineer Mikael Tollefsen, AIR-ICT
- PhD Candidate Brice Assimizele, AMO-Maritime Technology
In addition, several people will contribute towards the project, most importantly Associate Professor Johan Oppen at Molde University College, who is supervising the PhD candidate; and Senior Advisor of Maritime Safety at the NCA, Trond Ski.
Budget, Financing, and Timeframe
The project budget is 1 MNOK and is partially funded by AAUC (500,000 NOK) and Regionalt forskningsfond (RFF) Midt-Norge (500,000 NOK). The project intended startup date is 1 September 2013 and intended finish date is 30 June 2014. The project will, however, continue after this date funded by the professorship scholarship of the project manager (runs until
August 2015) and the PhD scholarship (runs until October 2016).
The main goal of the project is to develop new and stringent methods for dynamic resource allocation with maritime application based on methods from areas such as artificial intelligence, cybernetics, stochastic optimisation, informatics, and statistical physics.
The project manager has already developed a dynamic optimisation algorithm called receding horizon genetic algorithm (RHGA) able to assist VTS operators with dynamically optimising the positioning of the tug fleet. This work has resulted in two publications (Bye et al., 2010; Bye, 2012). Moreover, project group member Brice Assimizele has together with his supervisors developed an exact method for dynamic tug vessel allocation (Assimizele et al., 2013). A subgoal of the project is to further develop the RHGA with respect to implementation and use of real data and to continue the good cooperation with the NCA. It is also a subgoal to develop new methods from other scientific areas, and compare the methods. The methods shall be of use not only for the tug vessel preparedness of the NCA but should also be transferrable to other domains such as fleets of PSVs or other ships in the offshore industry. In addition, we wish to continue the work on developing exact methods (benchmarks) for further assessing such methods.
The work should result in presentations at international conferences and at least two publications on level 1 or 2. In subsequent projects, the goal is to refine and implement our methods for fleet optimisation as a standard decision-support tool for the NCA and for offshore companies.
The subgoals can be summarised as follows:
- M1 Further develop and improve existing RHGA algorithm for dynamic resource allocation of the NCA tugs, and explore other uses.
- M2 Develop new methods based on methods from the scientific areas listed under “Main Goal” that can be used across a number of disciplines and domains, including the NCA tugs but potentially also for PSVs or other ships in the offshore industry.
- M3 Develop benchmarks with exact solutions and theoretical computational speeds for quality assessment of the methods.
- M4 Present the work at international conferences and disseminate the results in two publications on level 1 or 2.
Problems in dynamic resource allocation are often complex, nonlinear, non-convex, stochastic, etc. In the field of operations resarch, the method of optimisation is commonly used to solve simpler problems statically, that is, by finding an optimal static solution for special cases that do not vary with time. However, the problem with this approach is the fact that the real worldis inherently dynamic and changing. The weather conditions are constantly changing, as is the traffic situation, thus new situations constantly arise and new choices must be made. In this project we want to combine the best parts from areas such as computational intelligence, cybernetics, operations research, stochastic optimisation, informatics, and statistical physics. We have already achieved promising results through the work of the project manager’s work on the TFO problem and respective publications, as well as the recent publication by the PhD candidate.
The main research problems that need to be solved in order to reach the subgoals M1–M4 can be summarised as follows:
- P1 Review factors influencing the VTS operators’ choices for positioning their fleet of tugs and implement these realistically in the RHGA algorithm. Examine the needs for optimisation of offshore vessels such as PSVs and adapt the RHGA to these needs. Perform more testing with realistic data and examine performance with respect to real-time
- P2 In relation to P1, examine alternative problem formulations and solutions
for control of the NCA tugs or offshore PSVs.
- P3 In relation to P1 and P2, define og possibly redefine problem formulations such that exact solutions exist and can be found in reasonable time. Develop test cases for these formulations that can act as benchmarks for comparing algorithms with respect to computational speed, performance, and more.
- P4 Achieve sufficiently good results with new methods such that the work is a significant contribution to world knowledge and worthy of publication. Find good conferences with respect to themes and problem description where the work can be presented and published. Develop contact networks within the offshore industry for future potential co-
operation and real-world implementation of methods. Write applications for more funding.
The following institutions are partners in the project:
- Aalesund University College
- The Norwegian Coastal Administration
- Molde University College