Abstract: In order to reduce the burden of patients’ medication costs, improve the quality of clinical medication, fundamentally improve the ecological environment of the pharmaceutical industry, promote the transformation of the pharmaceutical industry from market-driven to innovation-driven, and promote the solution of deep-seated institutional problems in the field of medical service system, in 2018, with the approval of the Central Comprehensive Deepening Reform Commission, the state organized the implementation of centralized and volume procurement of drugs. In 2020, the scale of the third batch of national organized centralized drug procurement reached tens of billions of yuan, with a total of 189 enterprises participating in the bidding process. Among them, 125 enterprises were selected, and 191 drug product specifications were selected, with an average price reduction of 53%. How to utilize the data of natural disasters in various regions over the past two decades, based on the concept of big data, scientifically and reasonably laying out drug logistics centers and drug delivery routes, minimizing drug delivery risks, and safely and efficiently distributing multi enterprise, multi type, large batch, national, and high time efficient national centralized procurement drugs to demand cities, have become a new problem that drug logistics enterprises urgently need to solve. This article starts from the actual needs of the country and enterprises, and focuses on solving new problems that arise in reality. It provides a new solution for the multi type, large batch, national, and high time efficient drug distribution problem after the national drug centralized procurement. It further improves the scientific nature and safety of the drug logistics center location distribution path, and provides a scientific theoretical basis for drug logistics decision-making.
In order to improve the timeliness and safety of drug distribution, we aim to address the problem of multiple types, large quantities, nationwide, and high timeliness of drug distribution. This article is based on the idea of big data, and based on the data of natural disasters occurring in the location of the alternative drug logistics center in the past twenty years, the ratio of the number of major sudden natural disasters occurring in the location of the alternative drug logistics center in the past twenty years to the selected age range of 20 is used as the predicted probability of the area suffering from major natural disasters in the future. Based on the predicted probability of major natural disasters in the future in the location of alternative drug logistics centers, a dynamic uncertainty drug logistics multi center location path optimization model is constructed, taking into account drug distribution safety, distribution costs, environmental protection costs, time satisfaction, and real-time road conditions. Based on the characteristics of the studied problem and in order to improve the efficiency of the algorithm, the fuzzyC-means algorithm is designed to be simple, has a wide range of problem solving, and is easy to apply to computer implementation. However, it is easily affected by the initial solution and falls into local optima. Particle swarm optimization (PSO) has strong global search ability, has more opportunities to solve the global optimal solution, and is easy to implement, high in accuracy, and fast in convergence. Using the solution obtained by particle swarm optimization as the initial solution of FCM can improve the computational efficiency of the algorithm. This paper makes full use of the advantages of fuzzy C-means clustering algorithm (FCM), particle swarm optimization (PSO) and Tabu search algorithm (TS) to design a hybrid PSO-FCM-TS algorithm. Based on the results of the second batch of centralized procurement bidding for drugs implemented by the national organization, calculations are conducted using PSO-FCM-TS, AG, TS, and PSO algorithms. The experimental results show that this algorithm improves convergence speed and has strong stability compared to AG, TS, and PSO algorithms.
This article aims to demonstrate the effectiveness of the constructed dynamic uncertainty drug logistics multi center location path optimization model and the designed PSO-FCM-TS hybrid algorithm. Based on the road conditions data of Gaode Map software and the randomly selected results of the second batch of national centralized procurement bidding for drugs, the dynamic uncertainty drug logistics multi center location path optimization model and the PSO-FCM-TS hybrid algorithm are used, seeking the optimal overall cost of drug distribution for pharmaceutical enterprises, as well as a distribution plan with a lower risk of natural disasters for drug logistics centers in the future. The empirical results indicate that the model can effectively determine the multi-center location path optimization scheme for drug logistics, and the algorithm has high convergence and stability.
Due to the difficulty of drug quality control, strong delivery timeliness, and high technical content, the professional and technical level of drug delivery personnel will have an impact on drug delivery. This article does not study the impact of human factors on drug delivery. Therefore, the impact of the professional technical level of drug delivery personnel on drug delivery will be the next research topic of the author.

Key words: with quantity purchase, drug delivery, PSO-FCM-TS hybrid algorithm, Live traffic

摘要： 本文以国家药品带量集中采购为背景。为提高药品配送安全和时效,降低药品配送成本。基于大数据思想,根据备选药品物流中心所在地近20年遭受自然灾害的数据,构建备选药品物流中心未来遭受自然灾害的预测模型,在综合考虑药品配送安全性、配送成本、环保成本、时间满意度和实时路况下,构建动态不确定性-药品物流多中心选址-路径优化模型。根据所研究问题的特点,为提高算法的效能,本文充分利用模糊C均值聚类算法(FCM),粒子群算法(PSO)和禁忌搜索算法(TS)等各自优点,设计了PSO-FCM-TS混合算法。最后,根据国家药品集中带量采购招标结果数据,对模型和算法进行了验证、对比和分析,研究结论为药品物流企业决策提供了科学依据。

关键词: 带量采购, 药品配送, PSO-FCM-TS混合算法, 实时路况