Static scheduling optimization method for the circular monorail transportation system in hilly and mountainous areas

Abstract: Circular monorail transportation system has been widely applied in hilly mountain areas. It is a high demand to improve the operation efficiency of static tasks in recent years. This study aims at the matching and scheduling requirements between mission points with the transport needs and transporters. The maximum full load was also considered comprehensively to determine the practice scenario of one-way goods transportation. The task was in the form of transporting cargo boxes from several loading points to a single unloading point. A mathematical model was established with the "carpooling" combined with the processing at the mission points and task assignment of transporters. The solving process of the transporter schedule was divided into two stages. In the first stage, the heuristic rule algorithms using different priorities were proposed to solve the "carpooling" combined processing with the goal of the least number of loadings of transporters. Combination algorithms included prioritization by adjacent position, task volume, and random sequence. In the second stage, a genetic algorithm based on variable neighborhood search (GA_VNS) was employed to solve the task assignment of transporters with the goal of the least number of transporter jams. GA_VNS was a hybrid intelligent algorithm with GA as a framework with the variable neighborhood search strategy, particularly with both global and local optimization. The performance of GA_VNS was compared with the standard GA, variable neighborhood search algorithm, overall matching rule, and random-restart hill-climbing algorithm. Transporting situations were simulated corresponding to three kinds of task batches. Five control experiments were carried out for each batch to improve the reliability of the experimental data. Experiments proved that the selected algorithms were superior. The designed combination algorithm with the task volume as the priority was solved fastest and was independent of the task batch. The operation time was generally between 0.3 and 0.4 s, with the lowest time cost. The feasible solution in 80% of the experiments was optimal among the three combination algorithms, and the solution advantage was concentrated in the large batch experiment, indicating more suitable for solving large-scale transportation in real time than the rest. The number of loadings in the remaining experiments was also only one unit more than the optimal solution. The solution of the assignment was also facilitated for the second stage scheduling in the validity. Therefore, the best performance was achieved in the combination algorithm with the task volume as a priority. The GA_VNS was employed to optimize the solution in multiple directions. GA_VNS significantly improved the quality of standard GA solution, and the number of transporter jams was reduced by 33.3%-100%. The higher stability was found in large-batch transporting situations, compared with the single variable neighborhood search algorithm. The probability of the minimum number of transporter jams appearing increased by 10% and 40%, but the operation time was about 10 times longer. The validity was better than other types of algorithms, such as overall matching and random-restart hill climbing. The finding can provide an approximately optimal solution to the static scheduling of the circular monorail transportation system in hilly mountain areas