应急物资需求与配送问题研究

发布时间：2018-05-14 16:44

本文选题：应急物资 + 需求预测　；参考：《西安科技大学》2014年硕士论文

【摘要】：近年来自然灾害频发，给人们的生命财产安全和社会经济发展带来了极大威胁，建立高效快速的应急救援系统迫在眉睫。应急物资的及时有效配送是救援工作的重中之重，而对应急物资需求的准确预测是保证配送工作顺利进行的前提。本文围绕应急救援工作，，展开了应急物资需求预测和应急物资配送问题的研究。首先，利用间接的方法预测灾区应急物资需求量。BP神经网络算法是预测灾区人员伤亡时的热点算法，针对标准BP神经网络学习效率低、易陷入局部最优的缺点，运用附加动量和自适应调整学习率相结合的方法对其进行改进；建立了基于BP神经网络的灾区人员伤亡预测模型，并对15次大型地震中的伤亡人数进行预测，验证了预测模型的合理性和算法的有效性；之后，在建立应急物资需求量预测模型的基础上，估算汶川地震中帐篷的需求量。其次，基于应急物资配送问题的特点，构建了以满足灾区需求为前提，总配送时间最短的应急物资配送问题的模型。最后，设计了求解应急物资配送问题的改进遗传算法，提出一种改进的比例选择算子，既保证了种群的多样性，又保证种群总是向着最优解靠近；融入具有较强局部搜索能力的模拟退火算法，在一定程度上克服了基本遗传算法的“早熟”收敛和易陷入局部最优的缺陷；利用TSP问题验证了改进的遗传算法的高效性；并以一次应急物资配送实例，验证了配送模型的合理性和改进的遗传算法的有效性。
[Abstract]:In recent years, the frequent occurrence of natural disasters has brought great threat to the safety of people's life and property and the development of social economy. It is urgent to establish an efficient and rapid emergency rescue system. The timely and effective distribution of emergency materials is the most important part of the rescue work, and the accurate prediction of the demand for emergency materials is the prerequisite to ensure the smooth distribution. In this paper, the emergency material demand prediction and emergency material distribution are studied around the emergency rescue work. First of all, using indirect method to predict the demand of emergency materials in disaster area. BP neural network algorithm is a hot algorithm when predicting casualties in disaster area. In view of the disadvantage of low learning efficiency of standard BP neural network, it is easy to fall into local optimum. The method of combining additional momentum with adaptive adjusted learning rate is used to improve it, and a prediction model of casualties in disaster area based on BP neural network is established, and the number of casualties in 15 large earthquakes is predicted. The rationality of the prediction model and the validity of the algorithm are verified. Secondly, the tent demand in Wenchuan earthquake is estimated on the basis of the establishment of the emergency material demand prediction model. Secondly, based on the characteristics of emergency material distribution, a model of emergency material distribution with the shortest total delivery time and the premise of meeting the needs of disaster areas is constructed. Finally, an improved genetic algorithm is designed to solve the emergency material distribution problem, and an improved proportional selection operator is proposed, which not only ensures the diversity of the population, but also ensures that the population is always close to the optimal solution. The simulated annealing algorithm, which has strong local search ability, overcomes the "premature" convergence of the basic genetic algorithm and is easy to fall into the local optimum to some extent, and verifies the efficiency of the improved genetic algorithm by using the TSP problem. The rationality of the distribution model and the effectiveness of the improved genetic algorithm are verified by an emergency material distribution example.
【学位授予单位】：西安科技大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：U116;F252

【参考文献】