A HYBRID ESTIMATION OF DISTRIBUTION ALGORITHM FOR ACONTAINER PRE-MARSHALING PROCESS

摘要

In a yard storage area, pre-marshaling means relocating the export containers into a proper arrangement in order to increasethe efficiency of the load process. The container pre-marshaling process aims to identify the best sequence of movements ofcontainer movements on an initial layout of the yard storage area, so that it reaches a desired final layout and satisfiesoperational constraints, while minimizing the total number of movements required to achieve the best sequence. Contrary tocurrent research, a new evolutionary algorithm is proposed and developed to solve the pre-marshaling problem and simulatethe solution. Our approach combines the key advantages of both evolutionary algorithms and the Mallows model. TheMallows distribution is used to model the pre-marshaling scenario, while an evolutionary algorithm is used to guide theoverall search process to identify the best performing sequences. The approach makes use of the Mallows model to describethe distribution of the solution space. The proposed algorithm is able to identify the next most probable movement in the yardstorage area. General and standard benchmarking and real-world cases served as input and test parameters in order to showthe performance of the proposed algorithm.