基于改进模拟植物生长算法的双目标逆向物流选址研究

发布时间：2017-12-27 06:24

本文关键词：基于改进模拟植物生长算法的双目标逆向物流选址研究　出处：《中北大学》2017年硕士论文　论文类型：学位论文

【摘要】：随着人类科学技术水平的提高,物质文明文化得到飞速的发展,但与之相伴随的就是资源的匮乏和能源的消耗殆尽,并且其中大部分的资源和能源是不可再生的。在这种处于危急关头的大背景下,如何将有限的资源和能源重新回收再利用,减小对于资源和能源的消耗是人们十分关注的问题之一。逆向物流选址在解决资源和能源消耗、实现可持续发展的目标中的作用日益显著。本文对模拟植物生长算法进行详细介绍后,对模拟植物生长算法存在的问题——对初始值的过度依赖性与得到最优解后无法迅速判断并自行终止问题进行详细剖析,并针对这些问题提出相应的改进方法,即在寻优过程中有效择优判断优秀的可生长点、将原算法的固定搜索步长进行智能化改变以及利用聚类法缓解算法对初始生长点取值的依赖性,详细解释了三种改进方法,并对改进后的模拟植物生长算法步骤及流程进行定义及解释说明。最后引入算例分情况讨论了改进方法并验证其有效性。证明改进算法采用的择优方法、搜索步长的变化以及利用聚类方法设置初始生长点均有助于解决模拟植物生长算法存在的优化效率较低、算法不能有效自行终止的局限性。在建立废弃物逆向物流选址模型时,本文考虑回收率确定、回收率不确定两种情况。在回收率确定情形下,假设废弃物如数回收,并采用斯坦福(Stanford)模型对废弃物产生量进行估算,考虑逆向物流选址设计中的经济成本与社会负成本,构建最小化经济费用和社会负效用的双目标函数模型,形成了一个综合考虑经济利益、社会利益的双目标整数线性规划,在建立模型后,利用改进模拟植物生长算法对选址模型进行求解。随后引入算例进行模型求解。考虑实际中存在回收率不确定的情形下在区域内构建废旧产品逆向物流选址的稳健模型,利用改进模拟植物生长算法对模型进行迭代求解最终得出经济成本、环境负效应最小的目标函数值以及回收处理中心的合适数量及位置。最后通过数值算例的引入验证了稳健模型在回收率不同的情况下的最优解和模型的稳健性。同时对回收率进行敏感度数值分析,得出随着回收率的增大双目标总成本增加,但是总体决策并没有发生改变。也就是说回收率的不同仅仅使目标函数最优值发生了变化,但最优的决策并没有发生改变,即选址模型决策中关于回收处理中心数量以及分配决策并没有随着回收率的增加而变化。
[Abstract]:With the improvement of the level of science and technology, material civilization and culture have been developing rapidly, but it is accompanied by the scarcity of resources and the depletion of energy, and most of them are non renewable resources. Under such a background of crisis, how to reclaim and reuse limited resources and energy and reduce consumption of resources and energy is one of the issues that people are very concerned about. The location of reverse logistics is becoming more and more important in the goal of solving resources and energy consumption and achieving sustainable development. This paper introduces the algorithm for the simulation of plant growth, the plant growth simulation algorithm to the initial value problems, excessive dependence and the optimal solution can be obtained after the detailed analysis and judgment to quickly terminate itself, aiming at these problems put forward corresponding improving method, namely in the process of optimization and effective judgment of outstanding merit the growth of fixed point, the search step of the original algorithm for intelligent clustering algorithm and the use of change point to the initial growth to alleviate the dependence, a detailed explanation of the three methods, and the improved algorithm simulation of plant growth steps and process definition and explanation. Finally, an example is introduced to discuss the improved method and verify its effectiveness. It is proved that the improved algorithm adopts the preferred method, the change of search step size and the initial growth point set up by clustering method, which is helpful to solve the limitation of the simulation plant growth algorithm, which is low efficiency and the algorithm cannot effectively terminate itself. In the establishment of the waste reverse logistics location model, this paper considers two kinds of conditions: the recovery rate is determined and the recovery rate is uncertain. The recovery rate under a certain situation, such as waste recycling and the use of assumptions, Standford (Stanford) model to estimate the amount of waste, the location in the design of economic cost and social cost of negative reverse logistics, establish the double objective function model of economic and social cost minimization of negative utility, forming a double objective integer linear programming is a comprehensive considering the economic benefits, social benefits, in the model, using the improved plant growth simulation algorithm to solve location model. Then an example is introduced to solve the model. Considering the recovery rate in the condition of uncertainty to construct a stable model of waste products reverse logistics location in the region found in practice, using the improved plant growth simulation algorithm for model iteration to solve the economic costs and environmental negative effect of minimum value of objective function and recovery processing center the appropriate number and position finally. Finally, a numerical example is introduced to verify the optimal solution of the robust model and the robustness of the model in the case of different recovery. At the same time, the sensitivity of the recovery rate is analyzed, and the total cost of the double target increases with the increase of the recovery rate, but the overall decision has not changed. That is to say, the difference of the recovery rate only changes the optimal value of the objective function, but the optimal decision has not changed. That is, the number of recovery centers and the allocation decisions have not changed with the increase of the recovery rate.
【学位授予单位】：中北大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;F252

【参考文献】