由於大多數之都市交通建設具『公共財』的特性，因此一般多藉由政府編列預算辦理興建。然而目前政府財政困難，所能編列之經費有限，因此無法滿足大量的交通建設需求，加上交通建設間具有『排擠效應』，為使有限資源做最佳之配置，以免造成浪費，有必要建立一套完整嚴謹的交通建設投資評估模式。但目前現行之中小型交通建設投資評估模式，由於政治力介入與規劃評估不切實際，使得原有之投資評估模式流於形式，故本研究將先針對目前中小型交通建設之現況進行探討與分析，並針對其重大缺失進行修正，並且考慮交通投資計畫間之相關性，進而建立一符合現況並且能考量整體性規劃之中小型交通建設投資評估模式。此模式可定義為一0-1型背包問題(Knapsack Problem)，由於具NP-complete之特性，針對大型問題欲求解其最適解將非常困難。所以本研究回顧近年來之巨集啟發式解法，並綜合其求解策略來發展數個啟發式演算法以有效率求近似最佳解，使得地方政府能夠篩選出最佳之交通建設投資組合，以期在有限資源下能夠發揮最大之績效。為測試本研究模式與解法的績效，本研究根據過去的歷史資料將對於評估模式所需之資料作一合理假設，進行範例測試研究與分析，並以C電腦語言撰寫電腦演算法，在個人電腦上執行求解，初步測試結果良好。 Most of the urban traffic constructions involve the nature of public goods. Public sectors are not satisfied most demands of traffic construction under budget constraints. The traffic constructions are character of crowding effectiveness. Construct evaluating models may promote usage of resource efficiently and avoid wasting which is an important task. Since the political intervene and the planning cann’t satisfy the real need, the existing evaluation model for the local traffic constructions is usually not practical. This study aims at analyzing and modifying a model for the local traffic constructions planning. This model considers the problem of correlation between alternatives. The model is formulated as a knapsack problem which is characterized as an NP-complete problem and is difficult to solve optimally for large-scale problems. This study reviews the recent heuristic algorithms and proposes some heuristic algorithms to solve the problem. The model could help the local government to choose the optimum investing combination of traffic constructions under budget constraints. In order to evaluate the model, we perform a case study concerning data from local governments in Taiwan. To enhance to applicability of the developed model and solution algorithms, we develop computer algorithms, using C computer language, and perform tests on personal computers. The preliminary results were impressive.
