Please use this identifier to cite or link to this item:
Measuring Efficiency of Fabless Integrated Circuit Design Houses by Using a Multiple Objective Programming Based Data Envelopment Analysis Approach
Integrated Circuit (IC) Design
Data Envelopment Analysis (DEA)
Malmquist Productivity Index (MPI)
Multiple Objectives Programming (MOP)
Efficiency Achievement Measure (EAM)
|Abstract:||效率衡量和提升公司績效一直是管理者所面臨最大的挑戰，公司在經營上的目標，如：追求卓越和長期發展，都和衡量公司的生產效率息息相關，管理者若能了解公司的經營績效，將對公司治理和公司價值的最大化有利。半導體產業是全世界最重要的產業之一且總體營收在西元2007年達到2535億美元，其產業價值鏈根據半導體設計和製造，後端製程則為半導體產品，半導體的供應鏈由四個部分組成，分別是IC設計，晶圓製造，封裝及測試和銷售。在2007年，IC設計的產值約為510億美元，為半導體產業總收入的20%，因此從半導體公司的經營者到投資人能夠了解IC設計公司的經營績效，將是重要的。過去，少有文章對全球的IC設計公司作績效評估，或者是著墨於傳統資料包絡分析法(Data Envelopment Analysis, DEA)的CCR和BCC模式，傳統的CCR和BCC模型評估各決策單位(Decision Making Unit, DMU)的經營效率，以各廠商的最適權重為基準，並不恰當。因此本論文的研究目的在於將DEA導入多目標規劃(Multiple Objectives Programming, MOP)模式，為全球IC設計公司求出一組公平、合理的權重分析，評估各家廠商的效率，依據文獻回顧和修正式德菲法由專家選出DEA的投入項和產出項指標，由效率達成衡量(Efficiency Achievement Measure, EAM)解出多目標DEA解答。藉由效率達成衡量，將能夠評估各IC設計公司的真正績效，並且找出各廠商的優勢和劣勢及提供經營策略，評估全球IC設計公司的經營績效可成為基礎產業分析和投資的基準，將多目標規劃導入DEA模型能成為績效評估的適當方法。|
The greatest challenge for managers of the world is to measure productivity efficiency and raise the productivity of firms. Measuring a firm’s productivity efficiency is related to its goal to pursue excellence and long-term development. Managers’ understanding of the efficiency measurement of a specific firm can help to maximize the value of corporate governance for both managers and shareholders. Semiconductor industry is one of the most important industries in the world, and the total revenue has already hit US$253.5 billion in 2007 while the industry structure or the value chain is based upon the design, manufacturing, backend processes of major semiconductor products. In general, the value chain of the semiconductor industry is consisting of four parts, IC design, wafer fabrication, device packaging& test (P&T) as well as marketing and sales. In 2007, the total revenue of the fabless IC design industry was about US$51 billion, which accounted for 20% of the whole semiconductor industry and played a significant role. Understanding the efficiency of the fabless IC design houses are critical for managers of the fabless IC design houses, managers of semiconductor foundries as well as personal investors. However, few literatures benchmarked performance of global fabless IC design houses. Further, the limited researches evaluated performance of global IC design houses introduced the traditional Data Envelopment Analysis (DEA) of CCR or BCC models, the performance evaluation models evaluating the performance of Decision Making Units (DMUs) by selecting their favorable weights, which can be misleading since the performance evaluation results were derived based on different bases of comparisons of DMUs. The traditional DEA models are not fair models from the aspect of improper weight derivations. Thus, the purpose of this paper to evaluate the efficiency of global leading fabless IC design houses by introducing a new and reasonable analytic model by introducing a Multiple Objectives Programming (MOP) based DEA method. The leading fabless IC design houses will serve as DMUs while the inputs and outputs of the MOP based DEA method will be selected by literature review and then refined by the modified Delphi method based on opinions of industry experts. Finally, the efficiency achievement measure (EAM) being derived by the MOP based DEA method. In this research, the real values as efficiencies of the world’s leading fabless IC design houses’ will be evaluated. In addition, according to the analyses based the MOP/EAM models, strength and weakness of the IC design houses can be demonstrated and strategies for enhancing the houses can be proposed. The performance evaluation results of global leading fabless IC design houses can serve as a basis for industry analysis, investment strategy, etc. In the future, the proposed MOP based DEA model can serve as an appropriate method for performance evaluations.
|Appears in Collections:||學位論文|
Files in This Item:
There are no files associated with this item.
Items in DSpace are protected by copyright, with all rights reserved, unless otherwise indicated.