Genetic Algorithm And Its Application In Stowage System Of Shipping Ships

In recent years Genetic Algorithm is paid more and more attention, and its application is getting more and more extensive, however there is less theoretical research on its convergence speed. Consequently, the theoretical analysis for the convergence speed of best individual preserved Genetic Algorithm is conducted in this paper, and on the basis of the analysis the traditional algorithm is optimized, which is to enhance the efficiency of the algorithm through enhancing the appearing probability of global optimal solution. In addition there is still an issue of premature convergence for Genetic Algorithm, so a method for overcoming premature convergence is proposed. This method successively runs Genetic Algorithm to obtain multiple different local optimal solutions in the process of searching optimal solution, then obtains global optimal solution from these solutions to be chosen. In this paper the issue of using Genetic Algorithm to learn the weight of neural network is also studied, and an optimized network architecture is proposed, which uses a network with multiple single outputs to replace a BP network with multiple outputs, so that features are simple in structure, and are more suitable to use Genetic Algorithm optimization. In addition, this optimized network architecture can be proved to have more superior optimal solution.0-1 programming is one of the most important integer linear programming. In the paper when Genetic Algorithm is adopted in a practical project of "Stowage System for Shipping Ships", available ships will be chosen to establish 0-1 programming model. In practical problems the number of available ships may be up to hundreds, so the method of all enumeration is obviously not suitable here. Therefore the idea of embranchment delimitation is introduced first in the paper so that the times of calculation are reduced dramatically, and the optimal solution can be obtained in a very short time in most cases. Then an improved Genetic Algorithm is proposed to solve the optimal solution against the situation that embranchment delimitation method can not obtain optimal solution in a shorter time.In fact it is difficult to solve such minority situations using any methods now known, and it is a better choice to use Genetic Algorithm obtaining a satisfactory solution. The experimental result has proved the effectiveness of the proposed method.