相關性運輸投資計畫選擇之研究—非模糊與模糊多目標規劃方法

Abstract

運輸投資計畫選擇 (TIPS) 問題，先天上具有多目標之本質，同時受到多項資源之限制。在未來不確定性與複雜性下，運輸投資規劃屬於不完整結構問題；因此，當進行運輸投資計畫選擇時，不僅要考慮多目標之本質，同時也要考慮模糊之特性。 運輸投資計畫之間，常具有相關性，而無法完全視為獨立性。本論文將運輸投資計畫區分為獨立類、互補類、替代類、以及同時互補替代類等四類型，再以相關領域之專家，應用非模糊與模糊方法，進行共識性之分類與相關程度之決定。 本論文探討TIPS問題時，區分為三種情況進行分析，即目標與資源均屬非模糊性、目標屬模糊性而資源為非模糊性、以及目標與資源均屬模糊性等。在達成多個目標與受到多項資源限制下，TIPS問題屬於0-1型多目標多次元背包問題，由於具NP完備性之特性，欲求取其最適解將非常困難。鑑此，本論文提出有效之啟發式解法……空間效率解法 (spatial efficiency algorithm)，以求取近似解，同時將此一解法加以推廣至具模糊性之TIPS問題。在模糊空間解法 (fuzzyspatial algorithm)中，需應用模糊數 (fuzzy numbers) 排序之方法；本論文應用Kim-Park排序法，並加以改良，使其不用決策者偏好加入，而又能同時考慮樂觀與悲觀之程度。 TIPS決策問題，除考慮目標達成之最大化外，對於資源之有效運用亦需同時考量，亦即期望可供使用資源之閒置情形為最小。本論文在運輸投資計畫之分類與相關程度之決定、TIPS問題三種不同狀況之求解、以及模糊數之排序等方面所提出之方法，均利用解釋例加以說明。最後，本論文提出未來可行之研究方向，以及在運輸投資規劃實務上可應用之決策問題。

Transportation investment project selection (TIPS) problem, are inherently multiobjective in nature and also subject to multiple resource costraints. Under a situation of uncertainty and complexity in the future, transportation investment planning issues are considered ill-structured. Therefore, when transportation investment planning is performed, not only the multiobjective nature but also the fuzzy characteristics have to be considered. Transportation investment projects are oftentimes characterized by interdependence, which could hardly be seen as independent. This dissertation will classify transportation investment projects into four categories, namely the independent, complementary, substitutive, and common complementary substitutive, and specialist from related expertise will be invited and nonfuzzy and fuzzy methodology will be utilized for deciding consensual classification and degree of interdependence. As this dissertation proceeds to investigate TIPS issues, three types of situations are respectively considered for analysis: objective and resource are both of nonfuzziness, objecive is of fuzziness and resource is of nonfuzziness, objective and resource are both of fuzziness. TIPS issues are of 0-1 multiobjective multidimensional knapsack problem under the condition of achieving multiobjective and being subject to multiple resource constraints, Its inherent trait of NP-completeness further intensifies its difficulty to find the optimal solution, Thus, this dissertation will put forward an effective heuristic algorithm, spatial efficiency algorithm, to obtain the approximate solution, and such algorithm will be extensively applied to TIPS issues of fuzziness. Fuzzy spatial algorithm should utilize the method of ranking fuzzy numbers, and this dissertation will as well apply Kirn-Park method and have it improved so that the preference of decision maker can be overlooked yet the degree of optimism and p-essimism will be taken into account. Aside from seeking the maximization of objective achievement, the decision making of TIPS issues will meanwhile consider the efficient utilization of resources and attempt to attain to minimum idle quantity of available resource. This dissertation will rely on illustrate examples for elucidation as it proceeds to methods presented to decide on the classification and degree of interdependence of transportation investment projets, to come out with solutions to the three situations of TIPS issues, and to rank fuzzy numbers. Finally, some issues for future research and some practical of decision problems on transportation investment planning will be discussed.