Genetic algorithm based hybrid optimization approach for container loading problem

Container loading problems, which are a vital part of global logistics and supply chain management, gain importance with operational complexity and cost effectiveness. In comprehensive operations where various products need to be delivered to numerous points worldwide, containers serve as the main transportation mechanism. These containers can come in various sizes and their weight capacities can vary significantly from each other. Loading and unloading of these containers effectively is of critical importance in terms of the efficiency and cost-effectiveness of the supply chain. Container loading issues are NP-hard problems considered in this context and included in the literature as a subgroup of cutting and packing problems. Operations research serves as an excellent tool for optimizing complex systems using mathematical models and algorithms to overcome these types of logistical challenges. Operations research and artificial intelligence share a common interest in solving complex problems with heuristic search. Artificial intelligence-based methods like genetic algorithm perform optimization processes through the intrinsic string representation of a particular system. The particular arrangement of this representation is a structure called a chromosome that has a decisive effect on the result of the optimization process. In this study, a hybrid solution is presented where the genetic algorithm optimization and the first-fit decreasing algorithm techniques are used together. Genetic algorithm optimization is an optimization technique that aims to find the most appropriate solution in the problem area by simulating the natural selection and evolution processes of the genetic algorithm. The first-fit decreasing algorithm is a powerful approach used in solving many optimization problems. This technique is based on the principle of not accepting solutions that yield a worse result than the current situation while searching for the best or most appropriate solution in the solution area. In the application of this hybrid solution, libraries like Geneticsharp were used to take advantage of the combination of genetic algorithm optimization and the first-fit decreasing algorithm. Geneticsharp is a library designed to support genetic algorithm solutions, frequently used in the literature, and with many advantages such as multi-threading. This library, with a wide user base and a mature API, is quite useful for genetic algorithm-based solutions. The visual aspect of the application was implemented using WebGL-based structures. WebGL is a web standard used for creating and processing 3D graphics, and in this study, it was used to visualize the operation and results of the algorithm. The effectiveness of this hybrid solution has been evaluated by comparison with other existing methods in the literature. The findings showed that this hybrid approach was not superior to existing methods in solving container loading problems. However, the findings also showed that this hybrid approach could hold a significant and meaningful place in the field of container loading optimization.