Optimization On Three-dimensional Loading In A Single Container

Container loading problem is an important sector in freight transportation. A reasonable loading scheme can improve the space utilization ratio and reduce the transportation cost, thus bring great economic profit. Theoretically speaking, container loading problem is a NP (Non-deterministic Polynomial, NP) complete problem, which is hard to find the optimal solution. Therefore, finding an effective algorithm to solve the container loading problem is quite meaningful to solving the practical problems of container loading and to calculating NP complete problem.This paper firstly gives a simple introduction about the container loading problem, secondly, analyzes and compares the heuristic algorithm (HA) and the genetic algorithm (GA) in detail, then proposes a hybrid genetic algorithm combined HA with GA. The algorithm uses the heuristic method to produce the initial solution group in the first place, and then decomposes the residual space by the three-space-decomposition method, uses the spatial merge principle to combine the free space with the usable space in order to achieve the full use of container, and cargoes are loaded with different order according to their types and numbers at the same time. In the end, the paper analyzes the GA's parameters; finds the best parameters' values which are fitting for this kind of question; and improves the crossover operation and mutate operation. In addition, a basic loading system is developed using VC++ and OpenGL which can produce a loading scheme for refrence.This paper mainly solves the weakly heterogeneous container loading problem, and also makes a simple research and discussion about the strong heterogeneous container loading problem as well as the problem which has the special constraints. Through the analysis and comparation with the examples, this algorithm puts up a good performance in solving container loading problem, and the final loading results have a great improvement. So applying this Hybrid Genetic Algorithm for solving the loading problem with a single container is very effective.