| Summary: | 碩士 === 中原大學 === 工業與系統工程研究所 === 105 === This research developed a heuristic algorithm to solve the optimization problems with soft-and-hard constraints in order to improve the solution quality and reduce the solution time. This study combined the information of soft and hard constraints with the decision tree method to generate an initial feasible solution, thereby reducing the generated initial feasible solution’s time. Furthermore, this study applied the local search concept and the information of soft and hard constraints to develop a new local search (greedy) method in order to obtain the better near-optimal solution and reduce the solution time.
In order to verify the feasibility of the proposed heuristic algorithm, this research used a radiologist scheduling problem as a case study. Through two numerical sets of different numbers of radiologists, this research tested the solution quality and solution time of two scenarios: the proposed heuristic algorithm with the particle swarm optimization (PSO) and the proposed heuristic algorithm with the bat algorithm (BA). The results showed that, for generating an initial solution, both PSO and BA methods with the designed initial solution mechanism could have better performance on solution time than both PSO and BA methods without the designed initial solution mechanism, and that for generating a local search, both PSO and BA methods with the greedy search could have better performance on solution quality than both PSO and BA methods without the greedy search.
|
|---|
