| Summary: | 碩士 === 國立高雄第一科技大學 === 資訊管理所 === 91 === This study developed a decision support system to assist the timetabling operation of a university environment. The one unique feature of this system is the constraints of both hard and soft nature constraint. Hard constraints are functional constraints that all feasible solutions must satisfy, and soft constraints are the preference type of requests that aim to improve a teaching environment. A scheduling solution must satisfy all hard constraints, but may not be capable of meeting all soft constraints. This study proposed a system development framework that is based on Constraint Satisfaction Problem, which allows user to investigate performances of proposed soft constraints before making final decisions. The solution framework is based on partial CSP, semiring-based CSP and hierarchical CSP, which treats the problem as a multi-objective decision problem. The iterative solution approach optimizes each objective sequentially and prevents de-generation of performances of earlier objectives. Backtracking with an arc consistency algorithm form the basis of the search algorithm for applying constraint propagation. Strategies of variable and value ordering played an important role in solution efficiency. The actual course scheduling needs of the Department of Information Management of National Kaohsiung First University of Science and Technology is used for the study.
|
|---|
