Product Development Process via Computer Simulation, DEA, DOE and Statistical Optimization Methodology

博士 === 逢甲大學 === 工業工程與系統管理學系 === 103 === Newness always wins. New product (NP) is a crucial part of a firm to survive, the ability of new product development (NPD) of a firm is essential to success. The efforts in earlier development stage create thicker value with low cost relative to total cost. Th...

Full description

Bibliographic Details
Main Authors: Francois Liang, 梁志鴻
Other Authors: Angus Jeang
Format: Others
Language:en_US
Published: 2015
Online Access:http://ndltd.ncl.edu.tw/handle/7k6y5x
Description
Summary:博士 === 逢甲大學 === 工業工程與系統管理學系 === 103 === Newness always wins. New product (NP) is a crucial part of a firm to survive, the ability of new product development (NPD) of a firm is essential to success. The efforts in earlier development stage create thicker value with low cost relative to total cost. Thick value is from the design alternatives creation and evaluation described in VA/VE Methodology. The value can be categorized into met the VOC, Time effect such as time to market, cost and quality consideration. Computer simulation applied to product development is the mega trend of NPD in 21th century. It combines with Data Envelope Analysis and statistical optimization which can evaluate the potential design alternatives. Then the potential one goes the detail design stage to gain customer’s satisfaction with short cycle of time to the market. This paper proposes a systematical approach integrated domain knowledge, computer simulation, parametrical and tolerance design, design of experiment DOE, and response surface methodology (RSM), optimization technology. The approach intents to solve the design problem which allow and integrate multi-pattern design introduced, design alternative creation and ranking, detail design with parameters and their tolerance simultaneously and optimization statistically. There is one example, which is complicated design problem in bicycle industry, implemented and gets a good result. The example is bicycle frame design with multi-static quality expected in the early design stage, which including 12 design pattern are introduced, there are 8 design alternatives exploring from each design pattern. Totally 96 design alternatives are introduced and the potential ones are selected according to different niche market (decision criteria). The potential one will go to the detail design, which is being of 16 design variables with 3 levels and two noise factors modeled and simulated by ANSYS. It is subject to the response of reliability, cost and dependability to optimize weight of bicycle frame.