基于神经网络的海工项目计划编制技术研究

发布时间：2018-03-18 02:18

本文选题：海洋工程项目　切入点：计划编制　出处：《哈尔滨工程大学》2016年硕士论文　论文类型：学位论文

【摘要】：海工项目计划是项目实施的依据,必须在项目开工之前编制完成,这样项目才能在进度计划的指导下高效有序的进行。通过调研发现国内的海工项目普遍存在计划无法指导执行而流于形式的问题,一个重要原因是海洋工程项目具有周期长,单件或小批量生产,EPC总承包合同,项目工作跨专业、跨部门甚至跨企业等特点,因此项目计划编制难度极大,往往项目已经开工而计划编制尚未完成。计划的编制可以看作一种决策问题,而人工神经网络在智能决策系统的应用中有良好的表现。为解决制约海工项目计划编制效率的关键问题,本文研究了基于人工神经网络技术的海工项目计划编制关键技术,开发了海工项目计划编制原型系统。本研究有利于我国海工企业计划管理的科学化和规范化。论文分析了国内外项目计划编制技术以及海工项目计划编制技术研究现状,阐述了人工神经网络技术的原理;深入研究了海工项目计划编制技术,并找出了制约海工项目计划编制效率的关键点;在此基础上,本文提出了两个需要解决的关键问题:工作分解结构(WBS, Work Breakdown Structure)分解和活动排序。针对WBS分解问题,建立海工项目工作分解结构模型,提出基于模块化神经网络的WBS分解方法;针对活动排序问题,建立海工项目作业排序模型,提出基于BP神经网络的作业排序方法。最后,将上述研究成果应用于实践,采用基于Spring的开发框架和B/S结构,设计并开发海工项目计划编制原型系统,验证方法的可行性和有效性。
[Abstract]:The sea engineering project plan is the basis for the project implementation and must be completed before the project starts. Only in this way can the project be carried out efficiently and orderly under the guidance of the progress plan. Through the investigation, it is found that there is generally a problem that the marine engineering projects in China are not able to guide the implementation and become a mere formality. One of the important reasons is that the offshore engineering projects have a long period of time. Single or small batch production EPC general contract, the project work cross-professional, cross-departmental or even cross-enterprise, so the project planning is very difficult, Often the project has been started and the plan has not been completed. The preparation of the plan can be seen as a decision-making issue. But artificial neural network has good performance in the application of intelligent decision system. In order to solve the key problem of restricting the efficiency of marine engineering project planning, this paper studies the key technology of marine engineering project planning based on artificial neural network technology. The prototype system of marine engineering project planning is developed. This study is beneficial to the scientific and standardized planning management of marine engineering enterprises in China. This paper analyzes the domestic and foreign project planning technology and the present research situation of marine engineering project planning technology. This paper expounds the principle of artificial neural network technology, deeply studies the planning technology of marine engineering projects, and finds out the key points that restrict the efficiency of marine engineering project planning. This paper presents two key problems to be solved: Work Breakdown structure decomposition and activity sequencing. To solve the problem of WBS decomposition, a WBS decomposition method based on modular neural network is proposed. Aiming at the scheduling problem of activities, a scheduling model for marine engineering projects is established, and a method of job scheduling based on BP neural network is proposed. Finally, the above research results are applied to practice, and the development framework based on Spring and the structure of B / S are adopted. Design and develop a prototype system for marine engineering project planning to verify the feasibility and effectiveness of the method.
【学位授予单位】：哈尔滨工程大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：P75

【参考文献】