Evolutionary Fleet Sizing in Static and Uncertain Environments with Shuttle Transportation Tasks-The Case Studies of Container TerminalsTri Thanh- Le
Shayan Kavakeb*, Trung Thanh Nguyen*, Zaili Yang and Ian Jenkinson
*Dept of Maritime and Mechanical Engineering, Liverpool John Moores University, Liverpool, United Kingdom
Published in: IEEE Computational Intelligence Magazine, Volume 11, Issue 1, pp. 55-69, Feb.2016(flagship journal of the IEEE Computational Intelligence Society, acceptance rate <10%, only accepted research with proven real-world applications).
This paper aims to identify the optimal number of vehicles in environments with shuttle transportation tasks. These environments are very common industrial settings where goods are transferred repeatedly between multiple machines by a fleet of vehicles. Typical examples of such environments are manufacturing factories, warehouses and container ports. One very important optimization problem in these environments is the fleet sizing problem. In real-world settings, this problem is highly complex and the optimal fleet size depends on many factors such as uncertainty in travel time of vehicles, the processing time of machines and size of the buffer of goods next to machines. These factors, however, have not been fully considered previously, leaving an important gap in the current research. This paper attempts to close this gap by taking into account the aforementioned factors. An evolutionary algorithm was proposed to solve this problem under static and uncertain situations. Two container ports were selected as case studies for this research. For the static cases, the state-of-the-art CPLEX solver was considered as the benchmark. Comparison results on realworld scenarios show that in the majority of cases the proposed algorithm outperforms CPLEX in terms of solvability and processing time. For the uncertain cases, a high-fidelity simulation model was considered as the benchmark. Comparison results on real-world scenarios with uncertainty show that in most cases the proposed algorithm could provide an accurate robust fleet size. These results also show that uncertainty can have a significant impact on the optimal fleet size.
Evolutionary computation, uncertainty, shuttle transportation, intelligent vehicles.