关键词: 物流, 网络, 海空协同, 优化模型, 文化基因算法

Abstract: As the South China Sea Islands are far from the mainland, many of the daily production and living materials need to be supplied from the mainland, and the seafood from the islands needs to be transported back to the mainland, so the establishment of an efficient logistics network is an important part of the construction of the South China Sea. In recent years, the Chinese government has expanded and built a few seaports and airports in the Xisha and Nansha Islands in the South China Sea to meet the production and living needs of the islanders. The construction of these transport facilities has laid a very favorable foundation for building a logistics network. In this context, the construction and optimization of a logistics network for sea-and-air collaborative transportation is of great theoretical significance and application value to further enhance the economic development capacity of the islands, strengthen the production capacity and quality of life of local people, and defend China's sovereignty over the South China Sea.
This paper establishes a logistics network optimization model with the minimum total logistics cost as the optimization goal, and comprehensively determines the site selection of marine center islands, the number and scale for construction berths, the types and quantity of transport ships and aircraft, shipping and air route configurations, storage capacity of warehouse and other issues, focusing on solving the location-inventory-route optimization problem of the logistics system under the collaborative transportation of sea and air transport modes. The model needs to optimize the logistics system for the daily needs of islanders under normal circumstances and needs to adapt to the air transport of emergency supplies under special circumstances such as typhoons. Based on the characteristics of the proposed model, we propose a memetic algorithm (MA), which combines a variable neighborhood search (VNS). Besides, some improvements are made in the selection of local search chromosome, the adaptive selection of neighborhood selection and the incorporation of memory search into fitness function calculation, which improves the efficiency of the algorithm. Finally, the effectiveness of the model and algorithm of this paper is verified by taking the optimization of the logistics network of Xisha and Nansha Islands in the South China Sea in China as an example. Four sets of different scale cases are compared using the MA-VNS algorithm and the genetic algorithm. The results show that the MA-VNS algorithm reduces the total cost by 5.50%-13.83%, the computation time by 5.88%-16.55% and the standard deviation by 15.28%-60.38% compared to the genetic algorithm, which indicates that the MA-VNS algorithm outperforms the genetic algorithm in terms of solution quality, efficiency, and stability for problems of different scales.
The research results can provide new ideas and methods for the construction of sea-and-air collaborative logistics networks in remote islands, which can provide technical support for the construction of sea-and-air collaborative transportation logistics networks in China's South China Sea Islands and have important theoretical significance and application value for strengthening the construction of the South China. This study only considers air transportation from a mainland airport to remote islands with airports. However, helicopters can also be used to transport supplies to islands. In such a situation, ships and helicopters can be used simultaneously, which may introduce greater complexity. This can provide a direction for future research.

Key words: logistics, network, sea-and-air collaborative, optimization model, memetic algorithm