Rail Multi Vehicle Scheduling Method for Intermediate Depot of Steel Plant Based on Improved Genetic Algorithm

Abstract

Appropriate transportation network plays a significant role in the economic development of country. The rising transport demand increases the congestion in railway networks, and thus, they become more interdependent and more complex to operate. A ground railway system is established in the reservoir area of steel plants to ensure smooth running of multi vehicle logistics transportation. In recent years, the development of the ground rail multi vehicle logistics system in the reservoir area of steel plants has been increased. However, such logistic systems need an intelligent dispatching control system to avoid the possibility of safety hazards, to follow optimal tracks and to effectively manage logistics operation. In this paper, an improved hybrid genetic algorithm is proposed to realize the decision making and control of multi vehicle scheduling. Task assignment to multiple vehicles and multi-stage subparent vehicle scheduling is performed based on the information obtained from the concerned subsection. The self-learning hybrid algorithm works on the data extracted from an improved population and suggests an optimized solution. The individual behavior and optimization process are updated by self-learning, which ensures the effectiveness of iterative evolution. The selection, cross mutation, and self-learning expert base operation methods of the hybrid genetic algorithm are optimized. The proposed system is evaluated by taking the multi vehicle logistics system of cold rolling intermediate steel depot as a case study. The improved algorithm is implemented in MATLAB and is compared with the traditional genetic and particle swarm optimization algorithms. Results of the analysis prove that the hybrid genetic algorithm of self-learning knowledge expert base is effective in solving the multi vehicle logistics scheduling optimization problem. With the inclusion of the self-learning knowledge expert base genetic algorithm, the convergence trend of the proposed algorithm is enhanced; the traditional genetic algorithm converges after 140 iterations, while in the proposed algorithm, it is reduced to 100 iterations. The evaluation reveals that the proposed algorithm is speedier than the traditional genetic algorithm (GA) and particle swam optimization (PSO); the average solution time of the proposed algorithm is 119.1 while that of GA, PSO is 137.4, 131.4, respectively. The proposed algorithm is applicable to improve the operation and efficiency of the logistics transportation system. The approach is beneficial in intelligent control of metallurgical production. The proposed algorithm has practical significance to be followed in the development of intelligent metallurgical production and logistics transportation.