The Analysis And Estimation Method For Product Development Time

In present competitive environment, it is necessary for companies to analyze and optimize the product development time. However, there is somewhat lacking in systemic analytical methods for the time problems. This dissertation explores the analysis and estimation method for the product development time. Firstly, the influencing factors to the time of design activities are analyzed. A mapping method is proposed to extract product characteristics. Then the time estimating problem is discussed, and two intelligent methods based on fuzzy neural network (FNN) and support vector machine (SVM) are put forward. After the times of design activities are obtained, the time model of the whole product development process can be presented.The main content of this dissertation is introduced in detail as follows:1. Time factors of design activities are identified and an extraction method for product characteristics is described. The design times of activities are influenced by many factors, among which product characteristics are important parts. In order to find out product characteristics at the early stage of product development, a fuzzy measurable house of quality (FM-HOQ) model is established. This model is applied to measure and map characteristics from customer's technical demands, with the decomposition idea of quality function deployment (QFD). For customer's functional demands, a mapping pattern of"functions-principle-structure"is taken on.2. An intelligent estimation method for design time based on FNN is proposed. There is a kind of nonlinear mapping relationship between time factors and design time, and neural network can perform this mapping well. Because the data of time factors consist of crisp numerical information and fuzzy linguistic information, a new FNN model is presented to fuse data and realize the estimation of design time. This model makes use of fuzzy compositive evaluation to simplify structure.3. Aiming at the problem of finite samples in design time estimation, a new method based on SVM is introduced. FNN is a machine learning method adopted under the conditions of a great deal of samples, while the available pre-existing design cases in companies are often finite. To overcome this disadvantage,fuzzy regression theory is combined withν-SVM, and a kind of fuzzy SVM named Fν-SVM is proposed. Then, a new intelligent time estimation method and relevant parameter-choosing algorithm are put forward.4. For the overlapping and iteration between design activities in concurrent product development process, a time computing model with its corresponding optimization method is presented. The design activity groups are divided into coupling and non-coupling activity blocks based on the programmed design structure matrix (DSM). According to the direction of information flow, rework time of each design activity caused by overlapping is calculated in turn. Then the time model of product development is formulated. Given cost constraint, the problem of shortest development time is transformed into constrained optimization problem, and a corresponding algorithm for this problem is proposed.Finally, an analysis and estimation system is developed based on the above researches. Some basic features of the system are described in this dissertation.