【作者】 朱洪程；

【导师】 张连营；

【作者基本信息】 天津大学 ， 管理科学与工程， 2010， 硕士

【摘要】 工程项目多目标综合优化是工程项目管理领域的研究热点之一。以往针对工程项目多目标优化问题的文献大多是研究项目工期和项目直接成本之间的定性关系。近些年也有文献将工程质量目标加入目标模型,建立了以工期—成本—质量三大目标为基础的定量优化模型。但是至今很少有文献在建立工程目标模型时考虑到环境影响控制和安全管理这些很重要的目标。此外,传统的目标优化算法对优化函数有很多要求,解法适用范围较小,不适宜作为对多目标函数模型进行优化的通用算法。本文首先对微粒群算法、遗传算法和免疫算法进行了简要介绍,在此基础上通过将遗传和免疫算法中的交叉变异、记忆选择等思想引入微粒群算法,得到了改进的混合算法——遗传免疫微粒群算法,大大提高了原算法逃逸局部最优的能力。接下来,本文将交互式方法与遗传免疫微粒群算法结合,得到更适合求解工程多目标优化问题的交互式遗传免疫微粒群算法,这种方法使得定量决策信息和定性偏好信息都能够进入多目标优化和决策系统,安全管理等不易量化的目标也能得到考虑,工程管理目标得到更全面的优化。随后,文章用定性和定量分析相结合的手段分析了工程工期与成本、质量之间的关系,并对工程与环境影响的关系进行了定量化处理,建立了工程项目工期—成本—质量—环境多目标综合优化模型。最后,本文为一个实际工程项目建立了多目标优化模型,利用交互式遗传免疫微粒群算法对模型进行求解,优化结果实现了工程质量—成本—工期—环境—安全目标的均衡最优,从而验证了本文提出的交互式遗传免疫微粒群优化算法在工程项目多目标优化问题中的实用性和有效性。更多还原

【Abstract】 Multi-objective optimization of construction project has become a hot point in the area of project management. In the past study focused on the qualitative relationship between duration and direct cost of the project. In recent years some studied join the quality objectives of project into the multi-objective model and established a quantitative optimization mode based on duration-cost-quality. But very few of the study include the control about environmental influence and security management. In addition, the traditional mathematic optimization method has many requirements for objective functions. Its solution scope is small, so not suitable for multi-objective function model as a universal algorithm for optimization.Firstly, this dissertation introduces particle swarm optimization, genetic algorithm and immune algorithm briefly. On this basis and through introducing crossover, mutation, memory choice into particle swarm optimization, hybrid algorithm has been improved - the genetic immune particle swarm algorithm, which greatly improved the algorithm’s ability to escape local optimum. Next, this method combines interactive method with genetic immune PSO algorithm and develop suitable method for solving engineering problems, interactive multi-objective optimization genetic immune PSO algorithm, which makes the quantitative decision information and qualitative preference information can enter multi-objective optimization and decision-making systems, not easily quantifiable goals such as security management are taken into account, project management objectives are more comprehensive optimized. Subsequently, the article analyzed the relationship between duration, cost and quality of the project through combination of qualitative and quantitative analysis, quantified the relationship between engineering and environmental influence and established a Schedule-Cost-Quality-Environment multi-objective optimization model. Finally, I established a multi-objective optimization model for a practical project, using interactive genetic immune particle swarm algorithm to solve the model. Results achieved the best balance of the project quality-cost-schedule-environment-safety objective, which verified the practicality and effectiveness of the proposed interactive genetic immune particle swarm optimization in multi-objective optimization problems in projects.更多还原