An Adaptive Island Evolutionary Algorithm for the berth scheduling problem

• Published in: Memetic computing Vol. 12; no. 1; pp. 51 - 72
• Main Author: Dulebenets, Maxim A.
• Format: Journal Article
• Language: English
• Published: Berlin/Heidelberg Springer Berlin Heidelberg 01.03.2020; Springer Nature B.V
• Subjects: Adaptive algorithms; Applications of Mathematics; Artificial Intelligence; Bioinformatics; Cargo ships; Complex Systems; Congestion; Containers; Control; Engineering; Evolutionary algorithms; Genetic algorithms; Heuristic methods; Islands; Marine terminals; Mathematical and Computational Engineering; Mechatronics; Operators; Production scheduling; Regular Research Paper; Robotics; Scheduling; Transport buildings, stations and terminals; Island model; Migration; Berth scheduling problem; Parallel evolutionary algorithms; Adaptive mechanism; Maritime transportation; Optimization
• ISSN: 1865-9284; 1865-9292
• DOI: 10.1007/s12293-019-00292-3

Increasing volumes of the seaborne containerized trade put additional pressure on marine container terminal operators. Long congestion periods have been reported at certain marine container terminals due to inability of the infrastructure to serve the growing demand, increasing number of megaships, port disruptions, and other factors. In order to alleviate congestion and avoid potential cargo delivery delays to the end customers, marine container terminal operators have to enhance the efficiency of their operations. This study focuses on improving the seaside operations at marine container terminals. A new Adaptive Island Evolutionary Algorithm is proposed for the berth scheduling problem, aiming to minimize the total weighted service cost of vessels. The developed algorithm simultaneously executes separate Evolutionary Algorithms in parallel on its islands and exchanges individuals between the islands based on an adaptive mechanism, which allows more efficient exploration of the problem search space. A set of extensive computational experiments indicate that the optimality gaps of the Adaptive Island Evolutionary Algorithm do not exceed 1.93% for the considered small-size problem instances. Furthermore, the proposed solution algorithm was compared against the other state-of-the-art metaheuristic algorithms and exhibited statistically significant improvements in terms of the objective function values.