基于遗传模拟退火算法的三维离线装箱优化问题研究

发布时间：2018-04-09 21:08

本文选题：三维装箱优化　切入点：离线算法　出处：《吉林大学》2014年硕士论文

【摘要】：装箱优化问题广泛应用于工业制造、计算机技术以及日常生活中方面，对于三维离线装箱问题而言，其最主要的应用领域为货运轮渡船舱配载、货运列车车厢配载、集装箱装载、汽车车厢装载等方面，跟物流行业联系较为紧密，但是此类优化问题求解过程较为复杂。在物流配送活动中，装箱效果的好坏对于配送成本将产生直接的影响，随着物流行业的快速发展，学者们对于物流配送领域的箱体装载研究日益深入。但是就目前情况而言，学者的研究都只能以某种算法得出其近似解，所设计算法的优化结果以及时间性能都不够理想，而市面上大多数装箱软件操作界面的用户友好性较差，装载效果也有待进一步改进。因此，针对三维离线装箱优化问题，设计高效快速的算法具有重要的理论和现实意义，不仅可以解决物流配送领域的货物装载问题，，也可以为后续装箱优化软件的设计和开发以及其他相关问题的研究提供借鉴依据。本文以三维离线装箱问题为研究对象，利用遗传算法和模拟退火算法集成的思路对该问题进行求解，并编写程序代码在Matlab环境下进行实现。首先，本文对国内外学者关于三维装箱优化问题的研究现状进行了总结，在此基础上，确定了本文所研究的主要内容以及采用的技术方法，并对三维装箱问题的主要研究方法进行了详细阐述，包括遗传算法、启发式算法以及混合遗传算法的主要优缺点，另外，还对本文的主要技术方法之一——模拟退火算法进行了详细介绍。其次，本文针对三维离线装箱优化问题的具体特点，提出了装箱优化模型具体的目标函数和约束条件数值化处理方法，构建出最终的数学模型，后续以该模型为导向，确定了本文所设计算法的基本要素以及遗传模拟退火算法的集成方法，在该算法集成的框架下，利用Matlab进行算法代码编程和实现，并利用其他学者论文中的数据对算法的有效性进行验证。最后，本文以长春市某汽车公司为例，对该公司的供应商配送数据进行整理，选取五个供应商为研究对象，利用前文所设计的算法进行实际装载，并将最终优化结果与各个供应商现有装载数据进行对比分析，进一步论证了算法和程序代码的有效性。
[Abstract]:The packing optimization problem is widely used in industrial manufacture, computer technology and daily life. For the three-dimensional off-line packing problem, the main application fields are cargo ferry cabin stowage, freight train compartment stowage.Container loading, car loading and other aspects are closely related to logistics industry, but the process of solving this optimization problem is more complex.In the logistics distribution activities, the effect of packing will have a direct impact on the cost of distribution. With the rapid development of logistics industry, scholars on the field of logistics distribution box loading research is increasingly in-depth.However, as far as the present situation is concerned, scholars can only get their approximate solution by some algorithm. The optimization result and time performance of the designed algorithm are not ideal, and the user friendliness of the operating interface of most boxed software on the market is poor.The loading effect also needs to be further improved.Therefore, it is of great theoretical and practical significance to design an efficient and fast algorithm for the optimization of 3D off-line packing, which can not only solve the cargo loading problem in the field of logistics distribution.It can also provide reference for the design and development of optimization software and other related issues.In this paper, the problem of three dimensional off-line packing is studied. The genetic algorithm and simulated annealing algorithm are integrated to solve the problem, and program code is written to realize it in Matlab environment.First of all, this paper summarizes the domestic and foreign scholars' research status of three-dimensional packing optimization problem, and on this basis, determines the main contents of this study and the technical methods used in this paper.The main research methods of 3D packing problem are described in detail, including the main advantages and disadvantages of genetic algorithm, heuristic algorithm and hybrid genetic algorithm.The simulated annealing algorithm, one of the main techniques in this paper, is also introduced in detail.Secondly, according to the specific characteristics of the three-dimensional off-line packing optimization problem, this paper puts forward the concrete objective function and the numerical treatment method of the constraint conditions of the packing optimization model, and constructs the final mathematical model, which is followed by the model.The basic elements of the algorithm designed in this paper and the integration method of genetic simulated annealing algorithm are determined. Under the framework of the algorithm integration, the algorithm code is programmed and implemented by using Matlab.The validity of the algorithm is verified by using the data of other scholars.Finally, taking an automobile company in Changchun as an example, this paper collates the supplier distribution data of the company, selects five suppliers as the research object, and uses the algorithm designed above to carry out actual loading.Finally, the optimization results are compared with the existing loading data of each supplier, and the validity of the algorithm and program code is further demonstrated.
【学位授予单位】：吉林大学
【学位级别】：硕士
【学位授予年份】：2014
【分类号】：TP18;U116

【参考文献】