在這講求高效率的年代，不論所處的環境為何，都一直不斷的強調如何在有限資源下取得最大利益。為了讓資源分配達到最佳效益，本篇論文提出以計算式智慧的方法來解決兩個資源分配的問題，一個是空戰戰機接戰決策的問題，另一為手術室排程規劃的問題。 在數位化戰場時代，透過科技提供迅速且有效的協助，將成為未來致勝關鍵。因此本論文將研究空戰戰場上戰機接戰決策的應用，提供符合最佳優勢的接戰決策策略。在空戰戰機接戰決策問題中，透過取得的空戰態勢，利用多層感知機來分析戰機遭遇時間，並評估戰機接戰時的優勢與設計反制策略的評估函式。透過自我組織特徵映射圖最佳化演算法提供最佳的反制策略，輔助決策者分析與選擇，以提昇戰機接戰決策之效能。 醫院投注許多資金於手術醫療設備上，手術室對一間醫院營運而言，扮演著舉足輕重的角色，好的排程方式會讓手術室的使用效率變高，進而提昇醫院的營運效率。本論文針對手術室排程規劃，提出基因演算法與自我組織特徵映射圖最佳化演算法兩個最佳化演算法，以便從眾多手術室排程的方法中，找出一個讓使用效率達到最佳的排程的方法，增進手術室的使用效率。 In this efficient-oriented generation, making much of pursuing the best benefit with restrictive resources is very important. How to optimize an objective function under precedence restrictions is a very demanding and challenging problem. This study proposes a computational-intelligence-based approach for solving two kinds of resource allocation problems. The first kind resource allocation problem is the aircraft assignment tactic problem and the other one is the operating room scheduling problem. On modem air warfare, it is important for a commander to quickly and effectively make a proper assignment of our aircrafts to attack enemy aircrafts. The goal of the aircraft assignment tactic problem is to find a proper tactic to assign our aircrafts to meet head-on enemy aircrafts with the objective of maximizing the expected predominance. One important factor in the computation of the expected preponderance value is the prediction of the flying time needed by our aircraft to encounter an enemy aircraft. We train an artificial neural network to predict the flying time based on 11 important features (e.g., the velocities, directions, distance, etc). After the preponderance matrix has been generated, the SOM-based optimization (SOMO) algorithm is adopted to find an assignment which can maximize the expected preponderance value. Operating rooms play a decisive role on a hospital’s finance because there are usually a lot of expensive instruments in the rooms and a lot of resources are also needed there. Therefore, effective operating room scheduling can reduce costs and ensure efficient use of operating rooms while maintaining a good quality of care. The second part of the thesis is to develop an effective operating rooms scheduling algorithm. This study proposes a genetic-algorithm-based and a SOMO-based operating rooms scheduling algorithms to determine operation schedules which can fully improve the utility rate of operating rooms.
