【作者】 李强；

【导师】 张静；

【作者基本信息】 中国矿业大学 ， 技术经济及管理， 2008， 硕士

【摘要】 工程项目由若干工作组成,在项目的管理过程中,项目的主要目标——工期、成本、质量密切相关,任何一方发生变化都将会引起其它方面发生相应的变化,并直接或间接地对工程项目产生影响,因此对项目的各项工作进行集成管理优化显得尤为重要。微粒群算法是一种进化计算技术,源于对鸟群捕食行为的研究。与其他优化方法相比,微粒群优化算法的优势在于容易实现同时又有深刻的智能背景,既适合科学研究,又特别适合工程应用。该论文查询了近20年的国内外相关文献,以系统科学理论为基础,从集成管理的概念出发,首先,提出了工程项目集成管理优化的基本理论与方法,建立了工程项目集成管理优化的概念模型,该模型的求解属于国际上公认的NP—hard难题之一,针对此难题提出了利用基于优先权的编码技术和拓扑排序知识把概念模型转化为数学模型,使得模型的求解成为可能,同时也避开了模型求解时惩罚算子繁琐的设计问题。其次,在研究工程质量的量化方程时,运用柯布—道格拉斯回归模型描述工程质量与工程工期、成本间的定量关系。接着,引入了微粒群算法,定义微粒的含义,使用线性加权方法构造微粒适应值函数,解决了多目标间比较时产生偏序问题的难题,从而实现了多目标优化模型向单目标优化模型的转化。而后,提出同时具有惯性权重和限定因子参数的新版本微粒群算法,在论文附件中编制其matlab求解源程序,运用在以管道水平定向钻穿越工程为工程实例的集成管理优化模型中;经过反复地实验,给出了适合该工程实例的参数组合,微粒群算法程序在求解过程表现出了高效的搜索能力,获得了满意的优化结果。最后,着重讨论了在微粒群算法参数设计中微粒个体意识与集体意识的比较分析和微粒群种群规模与协同搜索能力的关系;在工程项目参数变化分析中,探索性地讨论了资源限量变化对集成管理优化最优结果的影响和微粒适应值函数中权重变化的设计以及集成管理优化模型的逆问题求解方法等难题。该论文有图13幅,表3个,参考文献65篇。更多还原

【Abstract】 Project consists of several works. In the process of a project management, the relationships among project three significant goals those are time cost and quality are high connected. Once any one among those goals was changed, it will have great influences on other one directly or indirectly. Therefore, it is very significant to manage all the works as integration after optimizing.Particle swarm optimization is evolutionary computation technique. It came of the research on bird flock preying behavior. Compared with other optimization algorithms, particle swarm optimization superiorities consist in achieved easily and having profound intellect background. Particle swarm optimization was not only suit for scientific research but also suit for project application.After querying recent 20 years related reference in and out of china, based on system science theory, started with the conception of integration management, Firstly, this thesis advanced basic theory and method of project integration management optimization, setup conceptive optimization model of project integration management which belongs to international legalized NP-hard problem, and then transformed conceptive model into mathematics model by using priority-based encoding and topological sort technique so that made solving the model possible and avoided designing the complex penalty function. Secondly, when it came to the research of project quality equation, this thesis described the quantity relationship among project quality time cost by using Cobb-Dauglas model. Thirdly, this thesis imported particle swarm optimization and defined the particle to solve partial order problem when compared particle fitness value by constructed particle fitness function in linear weight method. It achieved that multi-object model transformed into single-object model successfully. Fourthly, this thesis brought up a new version particle swarm optimization which both had inertia weight and constriction factor parameters. The author wrote matlab procedure in appendix and applied it in integration management optimization problem that took pipeline horizontal direction drilling project for example. After time and time experimenting, the author found out suit parameters combination for given instance project. Particle swarm optimization performed high efficient search capability in solving process and gained approving optimum results. In the end, this thesis put stress on analyzing particle individual consciousness versus collective consciousness and the relationship between particle swarm size and ability of co-operation searching in particle swarm optimization parameter design. In the process of analyzing project parameters change, this thesis discussed the influences of constrained resource quantity on integration management optimization result and linear weight coefficient selection in particle fitness function design. The author also suggested a method for the inverse model of integration management optimization problem.There are 13 graphs 3 tables and 65 references in this thesis.更多还原