Linear Penalty Relation in Genetic-Based Algorithms in Block Relocation Problem-Weights

摘要

In our previous work, we introduced the block relocation problem with weights (BRP-W), in which a set of identically-sized items is to be retrieved from a set of last-in-first-out (LIFO) stacks in a specific order using the minimum fuel consumption. BRP-W considers container weight in fuel consumption and horizontal and vertical moves of the containers. A new "global retrieval heuristic - GRH" that uses twelve parameters to quantify various preferences when moving individual containers was introduced and embedded in a genetic algorithm to find optimal values for parameters. A nonlinear penalty function was used in the GRH. In this research, we improved the previous methodology by adopting a more practical linear penalty function in the GRH. The results for both methodologies are effective in identifying near-optimal parameter settings.