基于随机权重优化的电缆线路工程造价评估
发布时间：2018-08-31 13:58
【摘要】：科学技术的迅速发展使得电力工程通过扩张规模来满足社会需求,电缆线路工程作为电力线路工程的重要工程之一,需求量不断扩大,随之而来工程投资不断在增加,给电缆线路工程投资方带来大量问题,传统的工程造价预算方法已无法满足实际工程需求,有必要研究新的科学方法对工程造价进行估算。近年随着智能算法的引进,工程造价评估方面出现了很多评估方法,研究较多方法有模糊数学、灰色关联度、神经网络、粒子群及支持想回归等理论技术。通过分析电缆线路历史工程的特点,得出是历史工程数据呈小样本特点,属于小样本数据的学习问题。支持向量回归对小样本学习具有理想效果,因此将采用支持向量回归模型对电缆线路工程造价评估,而实际应用中支持向量回归由于参数选择盲目性,计算精度和泛化能力有所欠缺,无法满足实际应用的要求,粒子群算法对参数调节具有理想的效果,将借助粒子群优化支持向量回归参数,同时通过权衡随机权重改进粒子群,提高粒子群调节参数的精度,增加评估模型稳定性和估算精度。结合电缆线路工程初步设计,分析电缆线路工程造价数据特点,建立评估指标体系;通过数据归一化对历史工程数据进行标准化,利用主成分分析法对评估指标进行特征提取,提高指标的合理性和有效性;结合支持向量回归和改进的粒子群算法,建立电缆线路工程造价评估模型并进行实例分析;实例结果及分析表明该模型可以有效地估算电缆线路工程造价,能够有效指导电缆线路工程新建工程造价评估。
[Abstract]:The rapid development of science and technology makes electric power engineering meet the needs of the society by expanding its scale. As one of the important projects of power line engineering, the demand of cable line project is expanding constantly, and the project investment is increasing. It has brought a lot of problems to the investors of cable line engineering, and the traditional method of engineering cost budget can no longer meet the actual engineering demand. It is necessary to study a new scientific method to estimate the project cost. In recent years, with the introduction of intelligent algorithms, there have been many evaluation methods in engineering cost assessment. Many methods have been studied, such as fuzzy mathematics, grey correlation degree, neural network, particle swarm optimization and support regression. By analyzing the characteristics of the historical engineering of cable lines, it is concluded that the historical engineering data is characterized by a small sample, which belongs to the learning problem of the small sample data. Support vector regression has ideal effect on small sample learning, so support vector regression model will be used to evaluate cable line project cost, but in practical application, support vector regression is blind because of parameter selection. The calculation accuracy and generalization ability are not enough to meet the requirements of practical application. Particle swarm optimization (PSO) has an ideal effect on parameter adjustment. Particle swarm optimization (PSO) is used to optimize the parameters of support vector regression and to improve PSO by weighing the random weights. The accuracy of particle swarm optimization parameters is improved, and the stability and estimation accuracy of the evaluation model are increased. Combining with the preliminary design of cable line engineering, this paper analyzes the characteristics of cable line engineering cost data, establishes the evaluation index system, standardizes the historical engineering data by data normalization, and extracts the evaluation index by principal component analysis. Combining support vector regression and improved particle swarm optimization (PSO), the cost evaluation model of cable line engineering is established and the example is analyzed. The example results and analysis show that the model can effectively estimate the cost of cable line construction and can effectively guide the evaluation of new construction cost of cable line project.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TM75
[Abstract]:The rapid development of science and technology makes electric power engineering meet the needs of the society by expanding its scale. As one of the important projects of power line engineering, the demand of cable line project is expanding constantly, and the project investment is increasing. It has brought a lot of problems to the investors of cable line engineering, and the traditional method of engineering cost budget can no longer meet the actual engineering demand. It is necessary to study a new scientific method to estimate the project cost. In recent years, with the introduction of intelligent algorithms, there have been many evaluation methods in engineering cost assessment. Many methods have been studied, such as fuzzy mathematics, grey correlation degree, neural network, particle swarm optimization and support regression. By analyzing the characteristics of the historical engineering of cable lines, it is concluded that the historical engineering data is characterized by a small sample, which belongs to the learning problem of the small sample data. Support vector regression has ideal effect on small sample learning, so support vector regression model will be used to evaluate cable line project cost, but in practical application, support vector regression is blind because of parameter selection. The calculation accuracy and generalization ability are not enough to meet the requirements of practical application. Particle swarm optimization (PSO) has an ideal effect on parameter adjustment. Particle swarm optimization (PSO) is used to optimize the parameters of support vector regression and to improve PSO by weighing the random weights. The accuracy of particle swarm optimization parameters is improved, and the stability and estimation accuracy of the evaluation model are increased. Combining with the preliminary design of cable line engineering, this paper analyzes the characteristics of cable line engineering cost data, establishes the evaluation index system, standardizes the historical engineering data by data normalization, and extracts the evaluation index by principal component analysis. Combining support vector regression and improved particle swarm optimization (PSO), the cost evaluation model of cable line engineering is established and the example is analyzed. The example results and analysis show that the model can effectively estimate the cost of cable line construction and can effectively guide the evaluation of new construction cost of cable line project.
【学位授予单位】：华北电力大学(北京)
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP18;TM75
【参考文献】
相关期刊论文 前10条
1 陈洁;侯凯;高晓彬;;输变电工程造价合理性评价方法研究[J];南方电网技术;2016年08期
2 章昀s，
本文编号：2215203
本文链接：https://www.wllwen.com/kejilunwen/zidonghuakongzhilunwen/2215203.html
