【作者】 兰淑丽；

【导师】 张秉森；

【作者基本信息】 青岛大学 ， 计算机软件与理论， 2004， 硕士

【摘要】 决策支持系统(DSS)是当前计算机应用的热点之一，它将人们的智能资源与计算机的功能相结合，以改进决策质量，是一种处理半结构化问题并为管理决策人服务的基于计算机的支持系统。DSS涉及到数学、计算机科学、统计学、运筹学、知识获取、数据可视化、高性能计算、专家系统、行为科学和系统理论等众多学科的相关知识，是一门交叉学科。 线性规划是运筹学的一个重要分支，也是决策支持系统静态最优化数学规划方法之一，它作为经营管理决策中的数学手段，在当代决策中的应用是非常广泛的。因为实际决策问题大多是具有多个目标函数的问题，所以在线性规划中多目标线性规划方法的应用更为广泛。本论文就是对几种广泛应用于实际决策问题的多目标线性规划算法的求解性能等参数进行了探讨和比较研究，并对三种算法进行了程序实现。 按照决策者对实际决策问题求解过程的参与程度的多少，多目标线性规划算法可以划分为交互式算法和非交互式算法两种。决策者的参与对于多目标线性规划问题的求解是非常重要的，尤其是表现在决策者的偏好信息的表述上。那么，面对实际决策问题的时候，决策者和分析人员应该选择哪一种多目标线性规划算法呢?是交互式的还是非交互式的呢?因为不同的算法具有不同的特点，所以不同的算法应该适合于不同的决策问题，在众多的多目标线性规划算法中，如何来选择一个合适的算法也应是问题求解过程中一个非常重要的环节。本论文通过对几种多目标线性规划算法的深入探讨和研究，从而分析得出一些关于具体的多目标线性规划算法适合具体的实际问题的规律和结论。 考虑到当前对多目标线性规划算法的研究热点是对模糊多目标线性规划问题的研究，所以在求解问题时采用的多目标线性规划算法在分类上应属于非交互式的多目标线性规划算法，以节约求解时间。更多还原

【Abstract】 Decision Support System is one of the hotspots of computer applications in recent years. It combines human’s intelligence with computer’s function and improves the effect of decision making. DSS is a computer-based support system, it can deal with ill-structured problems, and it serves decision makers. DSS is a cross-subject that relates to mathematics, computer science, statistics, operational research, KA (Knowledge Acquire), data visibility, high performance computation, ES (Expert System), behavior science, system theory, etc.Linear Programming (LP) is not only an important branch of operational research, but also one of the static optimization mathematical programming methods in DSS. As a mathematical measure, LP is applied to management decision making widely. The application of Multiple Objective Linear Programming (MOLP) is wider; because there are much more problems involved multiple objective functions. This thesis discusses and compares several methods of MOLP that have been broadly used in practical decision making problems, and program three methods of them.According to how decision makers (DM) participate in solving problems, the method of MOLP can be divided into two kinds: interactive and non-interactive method. DM’s participation is important to solving the MOLP problems, especially some preference information given from the DM. The DM and analyst have to select one from many MOLP methods when they solve some MOLP problems. So, which kind of methods should be chosen and how to select one from so many methods? Because different methods have different characteristics, different methods should apply to different practical problems. And so, how to choose one suitable method should be one of theimportant steps. This thesis discusses and studies several MOLP methods deeply, and then analyzes these methods and gives some rules and conclusion about what kinds of practical problems these methods adapted to respectively.Considering the hotspot of the research in MOLP is in the fuzzy MOLP, we should use non-interactive methods to save time when we solved the fuzzy MOLP problems.更多还原