复杂产品设计过程任务分解、排序和分配优化技术研究

发布时间：2019-07-08 15:00

【摘要】：企业之间的竞争随着经济的增长越来越激烈,企业的竞争表现为产品的竞争,如何开发出具有竞争力的产品且开发周期最短、成本最低成为企业重点关注问题。在开发过程中,产品的结构与功能随着科学技术的发展也变得越来越复杂,导致产品的设计过程越来越复杂,不确定因素多,采用科学技术合理管理规划产品设计过程对于企业生存至关重要。为此本文对复杂产品设计过程的任务分解、任务排序、任务分配建模及其优化技术进行了研究。针对任务分解,提出了一种基于粗糙集识别耦合任务集的方法,通过复杂产品结构树进行任务最小化分解,采用粗糙集处理任务关联评价数据的不确定性,识别聚类耦合任务,实现任务分解。针对复杂产品设计过程开发时间长、成本难以计算,提出基于任务重叠的复杂产品开发过程仿真方法。采用信息输入矩阵和信息输出矩阵来量化任务的顺序重叠和反馈重叠;对设计过程的迭代特性采用任务返工概率矩阵、返工影响矩阵、学习效应曲线及质量改进矩阵描述;在任务重叠和迭代情况下,提出时间、成本、质量的计算方法,并改进了产品开发过程仿真算法。最后,通过一个实例证明了该模型的有效性和实用性。针对任务排序优化,提出了基于遗传算法和仿真相结合的任务排序优化方法。首先,采用任务完成时间最短、任务完成质量最高作为任务优化的目标函数,建立控制成本的约束规则;然后通过遗传算法调用仿真程序,仿真获得目标值的方式,对任务排序优化进行求解。针对任务的分配,提出质量—成本—时间的多目标任务分配模型。首先采用仿真获得在重叠和迭代反馈情况下,单个任务的执行时间、成本和质量;然后采用人员效率矩阵、人员成本矩阵、人员执行质量矩阵描述设计人员设计能力及成本;针对设计任务属性,采用设计结构矩阵、任务工作量矩阵、任务质量权重矩阵来描述;提出质量—成本—时间多目标的任务分配模型和并行任务分配执行策略;最后采用基于蚁群算法的任务分配算法,并用实例验证算法的有效性。
文内图片：

图片说明：图１－ｌＰｅｒｔｉ网模型逡逑
[Abstract]:With the growth of economy, the competition among enterprises is becoming more and more intense, and the competition of enterprises is the competition of products. How to develop competitive products with the shortest development cycle and the lowest cost has become the focus of attention of enterprises. In the process of development, the structure and function of the product become more and more complex with the development of science and technology, which leads to the more and more complex product design process and many uncertain factors. It is very important for the survival of the enterprise to use science and technology to manage the product design process reasonably. In this paper, the task decomposition, task scheduling, task assignment modeling and optimization technology of complex product design process are studied. Aiming at task decomposition, a method of recognizing coupling task set based on rough set is proposed. The task is minimized and decomposed by complex product structure tree. Rough set is used to deal with the uncertainty of task association evaluation data, cluster coupling task is identified, and task decomposition is realized. Aiming at the long development time and difficult to calculate the cost of complex product design process, a simulation method of complex product development process based on task overlap is proposed. The information input matrix and information output matrix are used to quantify the sequence overlap and feedback overlap of tasks; the iterative characteristics of the design process are described by task rework probability matrix, rework influence matrix, learning effect curve and quality improvement matrix. Under the condition of task overlap and iteration, the calculation methods of time, cost and quality are proposed, and the simulation algorithm of product development process is improved. Finally, an example is given to prove the effectiveness and practicability of the model. For task scheduling optimization, a task scheduling optimization method based on genetic algorithm and simulation is proposed. Firstly, the shortest task completion time and the highest task completion quality are used as the objective function of task optimization, and the constraint rules to control the cost are established, and then the task scheduling optimization is solved by calling the simulation program through genetic algorithm to obtain the target value. Aiming at the assignment of tasks, a multi-objective task allocation model of quality, cost and time is proposed. Firstly, the execution time, cost and quality of a single task are obtained by simulation under the condition of overlap and iterative feedback, then the design ability and cost of the designer are described by using the personnel efficiency matrix, the personnel cost matrix and the personnel execution quality matrix, and the design structure matrix, the task workload matrix and the task quality weight matrix are used to describe the design task attributes. A quality-cost-time multi-objective task allocation model and parallel task allocation execution strategy are proposed. Finally, the task allocation algorithm based on ant colony algorithm is adopted, and an example is given to verify the effectiveness of the algorithm.
【学位授予单位】：西南交通大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TB472;TP18

【参考文献】