行政院國家科學委員會專題研究計畫 成果報告
台灣與大陸線上遊戲產業策略聯盟之研究
研究成果報告(精簡版)
計 畫 類 別 ： 個別型 計 畫 編 號 ： NSC 96-2221-E-151-007- 執 行 期 間 ： 96 年 08 月 01 日至 97 年 07 月 31 日 執 行 單 位 ： 國立高雄應用科技大學工業工程管理系 計 畫 主 持 人 ： 王嘉男 計畫參與人員： 碩士班研究生-兼任助理人員：李光裕 碩士班研究生-兼任助理人員：黃南翔 報 告 附 件 ： 出席國際會議研究心得報告及發表論文 處 理 方 式 ： 本計畫涉及專利或其他智慧財產權，1 年後可公開查詢中 華 民 國 97 年 07 月 28 日
行政院國家科學委員會補助專題研究計畫
■ 成 果 報 告
□期中進度報告
台灣與大陸線上遊戲產業策略聯盟之研究
計畫類別：
■
個別型計畫
□ 整合型計畫
計畫編號：NSC 96－2221－E－151－007－
執行期間：
96
年
8 月 1 日至
97
年
7
月
31 日
計畫主持人：王嘉男
共同主持人：
計畫參與人員： 李光裕、黃南翔
成果報告類型(依經費核定清單規定繳交)：
■
精簡報告
□完整報告
本成果報告包括以下應繳交之附件：
□赴國外出差或研習心得報告一份
□赴大陸地區出差或研習心得報告一份
■
出席國際學術會議心得報告及發表之論文各一份
□國際合作研究計畫國外研究報告書一份
處理方式：除產學合作研究計畫、提升產業技術及人才培育研究計畫、
列管計畫及下列情形者外，得立即公開查詢
□涉及專利或其他智慧財產權，
■
一年□二年後可公開查詢
執行單位：
國立高雄應用科技大學，工業工程與管系
中
華
民
國
97
年
7
月
29
日
中文摘要
現今的遊戲軟體產業隨著網際網路的迅速發展且國人對國際化概念普及，在線上遊戲 這塊領域中，競爭非常激烈，所以選擇適當的聯盟伙伴，已成為重要課題。本研究以資料 包絡法(Data Envelopment Analysis)為基礎利用，結合啟發式演算法，找出最佳聯盟組合的 方法。並以台灣與中國多家線上遊戲上線公司來實際資料進行分析，由結果得知，本研究 能協助業界在線上遊戲產業中尋找最佳的聯盟結合。
Abstract
As the gradually competition market of online game, many companies are finding new opportunities in strategy alliance for enhance competition and profits and this has become an important topic in this industry. Based on data envelopment analysis (DEA) and heuristic technique, this study proposes a new approach with a verification technique to resolve the issues. The objective of this paper is to provide an effective approach to find the right strategic partner when an online corporation implementing strategic alliance, as well as the analysis of efficiency after alliance. Realistic data of 10 companies are collected from published stock market of China and Taiwan. Moreover, the proposed verification is used to adjust the final candidate results. This method can support top managers finding the best candidates of strategic alliance.
1. Introduction
Theestimated marketsalesofTaiwan’sgaming industry are39.1 billion NT dollars,showed by Topology Research Institute [1]. Its growth rate will rise 58% and 60.9% for Taiwan and China, respective. For the rapid growth of online game, many Taiwan’s gaming companies target China as their biggest market chance. Investment in China market has become the one of the major trends. The purpose of this paper is to develop an effective approach for Taiwan online companies to find the best candidate of alliance in China. Modified by the research of Wang et al [10], based on data envelopment analysis (DEA) and heuristic technique, this study proposes a new systematic approach with a verification technique to resolve the issues. The objective of this research is to provide an effective approach to find the right strategic partner when an online corporation implementing strategic alliance between two countries.
Charnes, Cooper and Rhodes (CCR) were first proposed model of DEA [2]. DEA itself is a non-parametric method for assessing the relative efficiency of decision making units (DMUs) based upon multi-inputs and multi-outputs. The primitive DEA model adopts the concept of production in microeconomics: efficiency = output / input. Banker, Charnes, and Cooper (BCC) developed a new model from the CCR model to understand the problems of pure technical efficiency (PTE) and scale efficiency (SE) [3].
Most of DEA applications are focused on performance evaluation [4], [5]. However, DEA can not only be used in performance evaluation and can be extended more. Some researches extend DMU to study strategy and merger [6], [7]. These studies try to build a model to find the best merger candidate. Wang et al use DEA to solve strategy alliance issue [7]. Wang and Sun [8] based on the fuzzy Lyapunov function for forecasting future issue. Cheng et al. [9] proposed a Wavelet-based Fuzzy Neural Network (WFNN) for classification and medical diagnosis for estimating unknown events. Nevertheless, few of above researches apply in technology oriented business and focus on candidate selection of strategic alliance and execute verification as well. The proposed approach is introduced in Section 2. Case study and result analysis are discussed in Section 3. Finally, concluding remarks and future directions are depicted in Section 4.
2. Proposed Approach
Performance change after alliance could be good synergetic effect, or deteriorated. To simplify this issue, following assumptions are made: (1) All enterprises are seeking to increase efficiency by decreasing inputs or increasing outputs. (2) Strategic alliance leads to the effective use of the resources. (3) The same kind of inputs or outputs can be added when strategic alliance is constructed.
Based on related papers and interviews with many managers in an industrial forum, the structure and details are demonstrated in Figure 1:
Figure 1. Procedures of proposed method
Step 1: Choose Input/Output Variables
1.1 Interview with industrial mangers, survey and literature review to find dominated inputs and outputs in this industry.
1.2 Use correlation analysis to analyze positive relationship between each variable of inputs and outputs. If the factors with the negative coefficients between inputs and outputs, they need to beremoved asfollowing DEA’sbasicassumption.
Step 2: Collect industry related data
2.1 Search online related enterprises to find all potential candidates to be our original DMU list. Five Taiwan companies and five China ones are been chosen.
2.2 Collect history data on candidate companies.
Step 3: Efficiency calculation of virtual alliances
3.1 Combining five Taiwan companies with five China ones in twenty-five new virtual alliances. When two companies are formed together to be a virtual alliance, we add the values of all variables, respectively. This is used to express that two DMUs are allied to be a new one. 3.2 Combine these twenty-five companies with the original DMU list and then calculate by
ranking as indifference curves (RIC). The detail of RIC will be dressed later.
Bad
Good
Change improper
variables
Step 4: Ranking change
checking
Step 1: Choose variables
Step 2: Collect related data
Step 3: Efficiency calculation
of all virtual alliances and all
DMUs
Recommendation
Recalculate each
virtual alliance with
other DMUs
Rank checking
Application of RIC
Step 5: Verification
Calculation
Interview with
Industrial Manager
Government
databases checking
Correlation
analysis
3.3 Evaluate and compare the efficiencies of the original companies with those of virtual ones.
Step 4: Checking changes of performance
Pick up those virtual alliances that efficiencies are better than its original companies.
Step 5: Verification Calculation
5.1 Choose one virtual alliance which picks up from Step4. Put it into original DMU list and remove those two companies which form this virtual alliance.
5.2 Recalculate by RIC. If the virtual alliance still reach the best rank level. We recommend it. Otherwise, reject it. Repeat from Step 5.1, till all virtual alliances picked by Step 4 have completed.
According to the characteristics of DEA, the score of DMU is equal to 1, which means the performance ofthisDMU isbest. When the score isnot1,then performance ranking can’t follow by score, because the score of different DMU might calculated from different reference sets [2], [10].
Hence, this paper uses a method which calls Ranking as Indifference Curves (RIC) to solve theDMU ranking issue. RIC usesthetheory of“IndifferenceCurves”theory from economics. From DEA theory, only DMU with best efficiency are located on the indifference curve. Therefore, the ranking can be distinguished by the indifference curves. This research tries to reorganize DMUs into several indifference curves. The usage is explained below.
First of all, execute the DEA calculation to find the DMUs of score 1, and all of them are set to be ranking 1. Afterwards, remove the DMUs of ranking 1. Execute the DEA calculation from the remaining DMUs. Find the DMUs of score 1 and set to be ranking 2. And so forth, till all DMUs have been ranked.
3. Case Study and Results Analysis
The case of this research is to help Taiwan’s gaming companies to find the best China candidate for alliance. This research chooses five major online game companies in Taiwan and other five in China as following: Chinesegamer, Soft-world, Gamania, Wayi, and Softstar in Taiwan; Haihong, Shjuyou Net-work, China Netease, Shanda, and Sohu in China.
Since CCR can provide the performance measurement of total operation efficiency, this research chooses CCR to accomplish efficiency assessment. After summarized by ChinaIDC’s [11] report, Market Observation Post System of Taiwan Stock Exchange [12], and previous research paper [7], this research chooses several related factors to be inputs and outputs. There are“salescost,”“operating income,”and “reserving surplus.” Alldataofrelated arecollected and shows in table 1 and the money unit is 10,000 New Taiwan Dollar (NTD). Moreover, this research used “marketefficiency,”asone of the input variables [10]. It is defined as the sum of the number of different types of game multiplied the profit ratio of market. The bigger the value is; the more company profit comes from online game. Table 1 shows the raw data of all DMUs.
Market efficiency Sales cost Operating income Reserving surplus Chinesegamer 450 52,911 452,306 240,803 Soft-world 1,305 2,176,160 2,706,286 954,756 Gamania 1,715 935,659 1,941,151 -291,201 Wayi 990 130,343 374,999 55,546 Softstar 745 124,123 479,790 173,440 Haihong 1,200 565,623 990,807 251,645 Shjuyou 150 796,291 1,083,166 119,806 Netease 225 339,441 1,718,400 -448,881 Shanda 2,115 1,600,643 4,411,878 151,658 Sohu 150 57,985 176,095 -19,098
When two companies are formed together to be a virtual alliance, we assume that the same variable of both companies can be summed together. This is used to express that two DMUs are allied to be a new one. According to the DEA characteristics, the values of all variable for any DMU can not be minus. However, the reserving surpluses of virtual DMUs still have minus values. Thus, this research adds a value of 1,000,000 to reserving surpluses of all DMUs to make them positive.
Combine5 Taiwan’scompanieswith 5 China’sresultsin twenty-five new virtual companies. Then, add original 10 companies to be total 35 DMUs. Following by our proposed calculation, the results are showed in Table 2.
Table 2. Ranking of DMUs and Virtual Alliances
RIC Rank DMUs and Virtual Alliances
1 Chinesegamer, Shjuyou, Netease, Sohu, Chinesegamer+Netease
2 Chinesegamer+Shjuyou, Chinesegamer+Sohu, Soft-world+Netease, Softstar+Netease 3 Soft-world, Softstar, Shanda, Soft-world+Shjuyou, Wayi+Netease, Softstar+Shjuyou 4 Netease, Chinesegamer+Shanda, Soft-world+Shanda, Soft-world+Sohu, Softstar+Sohu 5 Haihong, Chinesegamer+Haihong, Soft-world+Haihong, Gamania +Netease,
Wayi+Shjuyou, Wayi+Sohu, Softstar+Shanda
6 Wayi+Shanda, Gamania +Shjuyou, Gamania +Shanda, Softstar+Haihong, Gamania +Sohu, Gamania, Wayi+Haihong,Gamania +Haihong
company –the ranking of virtual alliance. Take Soft-world + China Netease for example, the original Soft-world is estimated as Rank 3 (see Table 2), but after alliance with China Netease, they become Rank 2 (also from Table 2). The“Difference”oflevelis3 –2 = 1. This means Soft-world can improve the operation efficiency by alliance with Netease. The greater the positive difference is; the better efficiency improvement of the alliance has.
Table 3. Ranking Comparison
Virtual Alliance Ranking
Taiwan Co. + China Co. Taiwan Company Virtual Alliance Difference
Chinesegamer+Netease 1 1 0 Chinesegamer+Shjuyou 1 2 -1 Chinesegamer+Sohu 1 2 -1 Soft-world+Netease 3 2 1 Softstar+Netease 3 2 1 Soft-world+Shjuyou 3 3 0 Wayi+Netease 4 3 1 Softstar+Shjuyou 3 3 0 Chinesegamer+Shanda 1 4 -3 Soft-world+Shanda 3 4 -1 Soft-world+Sohu 3 4 -1 Softstar+Sohu 3 4 -1 Chinesegamer+Haihong 1 5 -4 Soft-world+Haihong 3 5 -2 Gamania +Netease 6 5 1 Wayi+Shjuyou 4 5 -1 Wayi+Sohu 4 5 -1 Softstar+Shanda 3 5 -2 Wayi+Shanda 4 6 -2 Gamania +Shjuyou 6 6 0 Gamania +Shanda 6 6 0 Softstar+Haihong 2 6 -4 Gamania +Sohu 2 6 -4
Wayi+Haihong 2 6 -4
Gamania +Haihong 2 6 -4
From Table 3, four virtual companies have enhanced theirs ranking. They are Soft-world, Softstart, Wayi and Gamania. However, forming the virtual company will change the competition status in industry. We need verification to make sure final results. That’s why we need to execute the verification step. Here, we use one virtual company for example. We use Soft-world + Netease and other original DMUs (Chinesegamer, Gamania, Wayi, Softstar, Haihong, Shjuyou Net-work, Netease, Shanda, and Sohu) to recalculate RIC approach. The result is rank 2 and we record it. Then change to another virtual companies to recalculate, till all 4 ones have been gone through verification step. The results are summary as Table 4.
Table 4. Verification Recommendation Table
Virtual Comany RIC rank Results
Wayi+Netease 1 Recommended
Softstar+Netease 1 Recommended
Soft-world+Netease 2 No recommended
Gamania +Netease 2 No recommended
From Table 4, the ranks of Soft-world and Gamania have changed to be 2. Therefore, we change these 2 companies to be non-recommended status.
From Table 3 and Table 4, the results can be separated to 3 groups and discuss below. Group 1: Chinesegamer+Netease
Taiwan Chineseamer is 1st ranking and the virtual alliance, Chinesegamer+Netease, is still 1stranking. This alliance is good choice to keep the competition to be the leading edge.
Group 2: Wayi + Netease and Softstar + Netease
Obviously, these are changed to be rank 1. These two companies can get advantages from alliance with Netease.
Group 3: others
For others, they can not improve ranking, so there is no any advantage for alliance.
4. Concluding Remarks
Modified by research of Wang et al [7], we propose a new modified approach to find the better results. By our study, the results show that Chinesegamer, Wayi and Softstar allied with Netease would be the best combinations. This approach can be applied not only to estimate industrial alliance but also improve decision making for businesses.
past focus simply on alliance. The accomplishment of this study can lead to future research with some intangible factors in the field; (2) this research discusses the topic of alliance and provides a new possible evaluation tool when enterprises consider exploring alliance; (3) normally, all DEA related papers implement performance analysis focusing on “score.” However, we believe that under different circumstances, discussions made solely based on “score”can bemisleading. That is why we propose RIC.
Future direction will be considered involvement with intangible resources and the trend of inputs and outputs could be discussed as well.
5. Reference
[1] White handbook of software game industry, Topology Research Institute, 2006.
[2] Charnes,A.,W.W.Cooperand E.Rhodes,“Measuring theefficiency ofdecision making
units,”European Journal of Operational Research, Vol.2 No. 6, pp. 429-444, 1978.
[3] Banker,R.D.,A.Chanesand W.W.Cooper,“Somemodelsforestimating technicaland scale
inefficienciesin dataenvelopmentanalysis,”Management Science, Vol.30, pp. 1078-1092, 1984.
[4] Camanho,A.S.,and R.G.Dyson,“Costefficiency measurementwith priceuncertainty:a
DEA application to bank branch assessments,”European Journal of Operational Research, Vol. 161, pp. 432-446, 2005.
[5] Homburg, C., ”Using data envelopment analysis to benchmark activities,” International
Journal of Production Economics, Vol. 73, No. 1, pp. 51-58, 2001.
[6] Wang, C.N., and C.H. Wang, “A DEA application model for merger and acquisition in
high-tech business,”Proceedings of International Conference on Engineering Management, pp. 43-47, 2005.
[7] Wang, C., J. Tsai,K.Yang,“A StrategicAllianceApproach fortheonlineGameIndustry
between China and Taiwan,” Proceedings of International Conference on Innovative Computing, Information and Control, Sep. 5-7, Kumamoto, Japan, 2007.
[8] Wang, Y., and Z. Q. Sun, “Guaranteed cost control of discrete- Time fuzzy system based on
fuzzy lyapunov function approach,” International Journal of Computer Science and Engineering Systems, Vol. 1 No. 2, pp. 85-91, 2007.
[9] Cheng J. L., C. H. Chen, and C. Y. Lee, “Classification and Medical Diagnosis Using
Wavelet-based Fuzzy Neural Networks,”International Journal of Innovative Computing, Information and Control, Vol. 4. No. 3, pp.735-748, 2008.
[10] Norman, M., and B. Stocker., Data Envelopment Analysis–The Assessment of Performance,
John Wiley & Sons, Inc., 1991.
[11] China Software Game, IDC, 2006.
[12] Taiwan Stock Exchange Market Observation Post System. 2007; Available:
計畫成果自評
內容與原計畫完全相輔，並已收錄在The Third International Conference on Innovative Computing, Information and Control (ICICIC2008, EI) 的研討會論文集中。
對於學術研究、國家發展及其他應用方面之貢獻為開拓DEA 在策略聯盟上的應用，並 可應用到其他的產業領域上加以推廣，且提供公司高級主管在策略規劃上一套客觀的計量 工具，企業歷經合作後，實行策略管理的概念，經常發現採行策略管理時。 本研究主要的貢獻歸納如下：（1）過去以 DEA 為方法的文獻中多半為績效評估，本研 究針對高科技線上遊戲產業，使將要採取策略聯盟行動的廠商，能以 DEA 的工具來探討 策略聯盟；（2）在目前科技業市場的高度競爭下，對管理者評估策略聯盟來選擇對象時， 更符合其選擇之策略，並能利用此方向來提供改善對策，更為產業策略聯盟建構具備科學 根據及分析邏輯之架構；（3）過去以 DEA 探討企業合作之文獻為少數，本研究可讓後續的 研究有一個開端；（4）本研究能提供決策者評估策略聯盟選擇對象時，符合其選擇合作夥 伴之策略，並利用DEA 中差額變數分析的特性，來提供改善建議，為企業在進行策略聯盟 時，建構具備科學根據及邏輯分析之評估工具及架構；（5）且規劃運用不同於一般 DEA 文 獻中皆只針對Score 做績效評估，因此本研究預計使用的 Rank 排序法，比 Score 更能確切 得比較出受評單位的排序優劣。
對於參與之工作人員，獲得之訓練有：（1）DEA 計量方法的瞭解與操作；（2）在高度
競爭的科技業中，線上遊戲產業的認識與瞭解；（3）一套完整的問題研究與解決方式並且
可供推廣之研發成果資料表
□ 可申請專利 ■ 可技術移轉 日期：97 年 7 月 29 日國科會補助計畫
計畫名稱：台灣與大陸線上遊戲產業策略聯盟之研究 計畫主持人： 王嘉男 計畫編號：NSC 96-2221-E-151-007學門領域：服務系統與科技管理技術/創作名稱
線上遊戲產業策略聯盟之方法發明人/創作人
王嘉男 中文： 現今的遊戲軟體產業隨著網際網路的迅速發展且國人對國際化概 念普及，在線上遊戲這塊領域中，競爭非常激烈，所以選擇適當的 聯 盟 伙 伴 ， 已 成 為 重 要 課 題 。 本 研 究 以 資 料 包 絡 法 (Data Envelopment Analysis)為基礎利用，結合啟發式演算法，找出最 佳聯盟組合的方法。並以台灣與中國多家線上遊戲上線公司來實際 資料進行分析，由結果得知，本研究能協助業界在線上遊戲產業中 尋找最佳的聯盟結合。技術說明
英文：As the gradually competition market of online game, many companies are finding new opportunities in strategy alliance for enhance
competition and profits and this has become an important topic in this industry. Based on data envelopment analysis (DEA) and heuristic technique, this study proposes a new approach with a verification technique to resolve the issues. The objective of this paper is to provide an effective approach to find the right strategic partner when an online corporation implementing strategic alliance, as well as the analysis of efficiency after alliance. Realistic data of 10 companies are collected from published stock market of China and Taiwan. Moreover, the proposed verification is used to adjust the final
candidate results. This method can support top managers finding the best candidates of strategic alliance
