- 學位論文
建構知識導向之專案經驗模型於風險監控
Building a knowledge intensive experience model for project risk control
摘要
企業在專案的開發過程中，常面臨的有成本超支、時程延遲、需求變更、及新技術替換等影響等，加上專案的時限或資源的限制，容易成為影響專案成功的不確定因素。而風險是一項不確定的事件或狀況，若其發生則會引發至少一項專案目標，如時間、成本、範疇、品質等管理不當，可能會阻礙專案的進行，甚至直接影響專案的成敗。風險管理為專案管理之其中一部份，其目地是將專案中的不確定因素加以局限與管理，讓風險在專案生命週期裡皆可被加以管理與監控，本研究為了解決專案風險問題，以整合專案開發經驗為出發，藉由知識本體論可快速彈性整合邏輯和資料的特性，為了解決共享專案經驗的獨特性和複雜性以建構專案經驗管理模型，建立知識本體技術為基礎的知識庫系統，以如下步驟實作之：1.以知識工程方法步驟：透過領域界定、知識擷取、知識塑模及知識表達等程序建立軟體開發專案知識模型。2.定義風險因子所影響專案的範圍和程度，以表達風險識別、分析、規劃、監督於專案中的軟體風險管理流程。3.以推論規則實作，系統依據事實將歷史風險處理知識，提供給前端使用者做為解決問題的參考。本研究提供軟體專案經理人決策過程，於已知的風險點上提早預測警示並提供解決方案參考，預防決策過程可能會發生的錯誤流程和處理，同時專案裡的成員也可運用此模型，得知參考資訊至目前進行中的專案，並將新的專案現況回饋，成為專案內成員風險處理和經驗累積的系統來源。
並列摘要
In project development, enterprises may constantly encounter cost overrun, schedule delay, demand change, and update of new technology. These factors, along with project deadline or resource constraints, may easily become uncertainty factors that affect project success. Risk is an uncertain event or situation. Existence of a risk may affect management of at least one of project goals, including time, cost, scope, and quality, and impede smooth progress of the project and directly influence the success of the project. Risk management is a part of project management. It is aimed to confine and manage uncertainty factors in a project and make risks properly monitored during the project life cycle. To solve project risks, this study began with integration of project development experiences. Based on ontology’s characteristics of being able to quickly and flexibly integrate logics and data, an ontology-based knowledge base system was constructed to solve the uniqueness and complicatedness project experience sharing. The implementation procedure was as follows: (1) The knowledge engineering steps, including scope definition, knowledge acquisition, knowledge modeling, and knowledge representation, were followed to construct a software project development knowledge model. (2) The scope and degree of the impact of risk factors on the project were defined to represent risk identification, analysis, planning, and supervision processes in project management. (3) Based on inference rules and facts, the system provided past risk handling knowledge as a problem-solving reference for front-end users. The proposed system could provide software project managers an early alarm of a known risk point and a solution for reference, helping them effectively avoid incorrect processes and handling during decision making. Besides, project members could also use this model to obtain reference information and feed back the current status of the new project. Therefore, it could be a source of project risk handling knowledge and experience.