Multi-Objective Optimisation of the Generalized Bin Packing Problem

This paper explores the multi-objective optimization of the generalized bin packing problem, which involves allocating both compulsory and non-compulsory items into a set of bins. The items have attributes such as weight, width, height, and due date, while the bins have characteristics including capacity and cost. The primary objective is to minimize cost, typically by reducing the number of bins used. However, real-world scenarios often involve multiple, competing objectives, such as meeting item due dates and achieving load balancing. To address these complexities, the paper proposes a multi-objective evolutionary model to solve the generalized bin packing problem. This approach optimizes multiple objectives, such as minimizing cost and reducing item lateness, allowing decisionmakers to balance trade-offs between solutions represented as a Pareto front. The model was tested on one- and two-dimensional problem instances, demonstrating its ability to minimize the objectives while offering a range of conflicting solutions in certain cases. The results also identified some limitations of the algorithm, including premature convergence and insufficient solution diversity. The paper discusses potential reasons for these issues and suggests directions for future research to improve the algorithm's performance, ensuring better handling of multiple objectives and enhancing solution variety.