系統動態觀點之高科技產品擴散,替代與競爭模型之研究

Abstract

近年社會科學市場擴散的觀念引起相當廣泛的興趣，科技產品(包括設計服務及組 織)擴散之研究在行銷、產品創新以及經濟政策上的貢獻相當大。快閃記憶體(Flash Memory)是目前最重要的一種唯讀記憶體(ROM)，快閃記憶體多用於記憶卡以及聲音、影 像等資料的儲存。根據半導體產業之莫耳定律，晶片上之電晶體的數目每18 個月成長 一倍，成長理論分析快閃記憶體的擴散提供了市場動態不同的看法。然而，快閃記憶體 之動態成長至今尚未被探討。另一方面，從全球佔有率來看，台灣個人電腦在全球市場 具有相當大的重要性，由於筆記型電腦價格不斷地下降，消費者有更多選擇，促使桌上 型電腦使用者轉而購買筆記型電腦，因此筆記型電腦是否會取代桌上型電腦的議題遂受 到矚目。 本論文以快閃記憶體與個人電腦為實例，首先應用成長模型探討快閃記憶體之動態 成長，雖然Bass 的模型在創新擴散中被廣為研究，但模型的假設限制了它的應用性， 因此本研究首先嘗試放鬆Bass 假設，模型上考慮非齊一性、非對稱性以及潛在採用者 之異質擴散。其次本研究比較非線性最小平方法(NLS)與基因演算法(GA)來估計多世代 Bass 模型並根據最佳績效結果來預測快閃記憶體之未來市場。最後，乃應用生物生態學 中修正的Lotka-Volterra 的數學模型，以電腦模擬來分析台灣個人電腦產業之桌上型 電腦及筆記型電腦之動態競爭過程。研究結果顯示Bass 模型增加非齊一及異質性兩個 參數探討快閃記憶體的成長較能描述快閃記憶體的擴散，另外用非線性最小平方法(NLS) 與基因演算法(GA)兩種方法來評估多世代快閃記憶體的成長，結果顯示基因演算法(GA) 更能夠有效地估計多世代快閃記憶體的擴散。此外， 利用生物生態觀點的 Lotka-Volterra 模型來分析個人電腦的競爭關係，發現桌上型電腦(DT)與筆記型電腦 (NB)的關係呈現純粹競爭的模式。 總之，本研究已透過放鬆Bass 模型的假設來探討單代與多世代快閃記憶體的擴散， 並進而成功地運用Lotka-Volterra 模型分析了台灣個人電腦的動態競爭關係。

The study of the diffusion of high-technological products (including services and organizations) benefits the operation of marketing, innovation of products, and formulation of economic policy. Moore's Law in semiconductor industry causes the number of transistors on a chip to double every 18 months. Flash memory is one of the important non-volatile memory components and has been widely used in modern information-technology (IT) industry. The explosion of the consumer electronic market has been sustained by the availability of semiconductor non-volatile memory technology. However, the growth of the flash memory has not been adequately explored. Analysis of the diffusion of flash memory using the growth theory may provide insight into the dynamics of the IT market. Shipment of personal computers (PCs) in Taiwan similarly plays a crucial role in the global market. Desktop (DT) PCs users have greater incentives to switch to notebook (NB) PCs due to the low price of PCs. Competition issue is thus attracted if DTs will be replaced by NBs. In this thesis, we explore the marketing dynamics, substitution, and competition of flash memory and personal computers. First, we apply an extended diffusion model (i.e., the Bass model) to explore the marketing dynamics of NAND-type flash memory. Despite Bass’ general acceptance in the literature on the diffusion of innovations, nine assumptions of the model limit its applications. Our study considers the non-uniform influence, the asymmetric diffusion, and the heterogeneity of the population of potential adopters which are relaxing assumptions by the Bass model. We further employ the nonlinear least squares (NLS) and the genetic algorithm (GA) in estimation of a multi-generation of NAND flash memory, including forecasting of the market. Finally, based upon a biologically evolutionary argument, we investigate the innovative growth and competitive relationship of personal computer shipments in Taiwan. Using a revised Lotka-Volterra model, we mathematically analyze the v dynamic competitive relationship between shipments of DTs and NBs, where management implication is drawn. The results of this study have shown that the Bass model, with the aforementioned relaxed assumptions, well describes the diffusion of NAND flash memory since 2000. In addition, GA estimates the marketing growth of a multi-generation NAND flash memory more efficiently than does NLS. The revised Lotka-Volterra model exhibits a pure competition between DT and NB in the past two decades. To sum up, we have explored the diffusion of a single generation and a multi-generation of NAND flash memory by relaxing the assumption of Bass’ model, and have also successfully applied the revised Lotka-Volterra model to analyze the dynamic competitive relationship of Taiwan’s personal computers.