資訊管理學報
陳柏榮;陳怡璋;陳安斌;韓復華;
頁: 63-78
日期: 2007/10
摘要: 本研究應用限制規劃方法來建立投資組合模式，適當地納入技術面、總經面與股市連動關係的專家知識，使之能依據環境的變化，對投資組合操作策略做出適當之調整。此外，本研究以投資組合限制規劃模式為核心，開發出一套「智慧型投資組合決策輔助系統」，提供使用者彈性處理實務投資操作邏輯規則的機制。該系統可籍由技術指標與總經指標的狀況，動態調整風險承受程度、個股資金分配、現金部位，打破傳統投資組合模式靜態的操作方式。 本研究以民國90年台灣股票市場之歷史資料作測試，每半個月為一期共分成24個測試期，並以「大盤指數」與「投資標的股票指數」為績效比較之測試標竿。實證結果發現，利用投資組合限制規劃模式求解產生之投資組合在保守考慮2%的交易成本下之年報酬為62.4%，遠高於大盤指數的22.8%與投資標的股票指數的34.8%；若以夏普比率做基準，投資組合在24個測試期中有20期皆能打敗測試標竿。整體而言，投資組合具有在低風險下尋求高報酬之特性，證實本研究提出之創新概念確實能有效提升投資組合之績效。
關鍵字: 限制規劃;投資組合模式;專家知識;技術指標;總經指標;
Abstract: Traditional MP-based optimization portfolio models are difficult to incorporate many critical macroeconomic and technical factors in quick changing investment environment. This research took the portfolio selection as a constraint satisfaction problem (CSP), and developed a constraint programming model to take expert knowledge into considerations. In addition, a decision support system is also developed to provide users a friendly interface to do preference setting and obtain the portfolio solution after processing by ILOG OPL Studio. Stock market data of Taiwan in 2001 were applied to test the proposed portfolio constraint programming model. Taiwan Stock Weighted Index and target stock index are adopted as benchmark for comparison. Our proposed model yielded an annual return of 62.4%, which is much higher than that of Taiwan Stock Weighted Index, 22.8%, and is also higher than that of the target stock index, 34.8%. The results showed that the portfolio constraint programming model provides an effective planning tool for portfolio analysis.
Keywords: portfolio model;constraint satisfaction problem;constraint programming;expert knowledge;decision support system;
瀏覽次數: 15770 下載次數: 121
引用 導入Endnote
頁: 63-78
日期: 2007/10
摘要: 本研究應用限制規劃方法來建立投資組合模式，適當地納入技術面、總經面與股市連動關係的專家知識，使之能依據環境的變化，對投資組合操作策略做出適當之調整。此外，本研究以投資組合限制規劃模式為核心，開發出一套「智慧型投資組合決策輔助系統」，提供使用者彈性處理實務投資操作邏輯規則的機制。該系統可籍由技術指標與總經指標的狀況，動態調整風險承受程度、個股資金分配、現金部位，打破傳統投資組合模式靜態的操作方式。 本研究以民國90年台灣股票市場之歷史資料作測試，每半個月為一期共分成24個測試期，並以「大盤指數」與「投資標的股票指數」為績效比較之測試標竿。實證結果發現，利用投資組合限制規劃模式求解產生之投資組合在保守考慮2%的交易成本下之年報酬為62.4%，遠高於大盤指數的22.8%與投資標的股票指數的34.8%；若以夏普比率做基準，投資組合在24個測試期中有20期皆能打敗測試標竿。整體而言，投資組合具有在低風險下尋求高報酬之特性，證實本研究提出之創新概念確實能有效提升投資組合之績效。
關鍵字: 限制規劃;投資組合模式;專家知識;技術指標;總經指標;
Development of a Decision Support System for Portfolio Selection Based on Expert Knowledge: A Constraint Programming Approach
Abstract: Traditional MP-based optimization portfolio models are difficult to incorporate many critical macroeconomic and technical factors in quick changing investment environment. This research took the portfolio selection as a constraint satisfaction problem (CSP), and developed a constraint programming model to take expert knowledge into considerations. In addition, a decision support system is also developed to provide users a friendly interface to do preference setting and obtain the portfolio solution after processing by ILOG OPL Studio. Stock market data of Taiwan in 2001 were applied to test the proposed portfolio constraint programming model. Taiwan Stock Weighted Index and target stock index are adopted as benchmark for comparison. Our proposed model yielded an annual return of 62.4%, which is much higher than that of Taiwan Stock Weighted Index, 22.8%, and is also higher than that of the target stock index, 34.8%. The results showed that the portfolio constraint programming model provides an effective planning tool for portfolio analysis.
Keywords: portfolio model;constraint satisfaction problem;constraint programming;expert knowledge;decision support system;
瀏覽次數: 15770 下載次數: 121
引用 導入Endnote