三階複選目標規劃法於森林疏伐減碳規劃之研究

A Study on 3-level Multi-choice Goal Programming for Carbon Reduction by Forest Thinning

陳瑩達(Ying-Ta Chen)

指導教授 ： 鄭欽龍

國立臺灣大學/生物資源暨農學院/森林環境暨資源學研究所/博士(2011年)

https://doi.org/10.6342/NTU.2011.02567

摘要

對森林經營者而言，選取適當工具規劃可達成多目標森林經營的疏伐作業並非易事。本研究旨在改善複選目標規劃技術之求解效率並探討其相關技術於疏伐減碳等多目標森林經營之適用性。在確認疏伐作業的成本效能實屬可行後，本研究首先以二元搜尋演算法結合複選目標規劃法，提出三階複選目標規劃法改良複選規劃法之求解效率，並以可供檢證之加拿大多目標森林經營案例比較兩者在多目標森林經營之求解效率。接著，利用多段目標規劃法配合線性規劃法，同時將疏伐面積及疏伐程度作為決策變數，以此建置新竹林管處柳杉人工林疏伐減碳規劃模型並探討該方法的適用性。確認三種方法在林業經營規劃可行後，本研究整合三種方法提出新竹柳杉人工林之最適疏伐減碳多目標規劃。
由加拿大多目標森林經營案例之驗證結果可知，複選目標規劃法相較於目標規劃法可迅速協助森林經營者找出最適經營期程，但目標值設定偏差則可能無法找到較佳解。三階複選目標規劃法則可在擴增搜尋範圍後，以每次運算減少一半搜尋範圍的方式，避免純粹使用複選目標規劃程式求解時的設定與比較，以有效找到較佳解。多段目標規劃法配合線性規劃法建置之新竹林管處人工林疏伐減碳模型研究結果顯示，該方法可一次模擬768種疏伐組合並找出最適疏伐期程以達到最大化碳吸存量的目標。本研究最後整合三階複選目標規劃法及多段目標規劃法進行新竹林區管理處之疏伐減碳等多目標規劃。研究結果指出，森林經營者應用此二方法可避免設定多種疏伐制度及目標值組合的模擬與比較，迅速找到能達成碳吸存、就業機會及土壤保護經營目標之最適疏伐期程。

關鍵字

森林經營 ； 複選目標規劃 ； 森林疏伐 ； 碳吸存 ； 多段目標規劃

並列摘要

It is not an easy task for forest managers to choose proper techniques to plan an appropriate thinning schedule for multiple forest management. The purpose of this paper is to improve the efficiency of multi-choice goal programming (MCGP) in searching better solutions and discuss the application of its related techniques toward various forest management problems. After confirming the feasibility of thinning in by cost-effectiveness analysis, the 3-level MCGP is firstly proposed to complete the insufficiency of the MCGP technique by combining binary search algorithm with MCGP and verify its efficiency by a credible Canadian multiple forest management case. Second, the application of combining multi-segment goal programming (MSGP) with linear programming to take both thinned area and thinning intensity as decision variables is evaluated by building a model of reducing carbon by thinning for artificial Sugi. in Hsinchu Forest District. Finally, both three techniques is applied to propose a thinning schedule for Hsinchu Forest District to achieve multiple forest management when carbon reduction is considered.
The results show that MCGP can rapidly help forest managers to search for a proper planning horizon for Canadian multiple forest management case. But, forest managers may miss a better ideal solution while the set target goal values are far from the best ideal solution. After expending the search range, 3-level MCGP can reduce the search range by half in each run and avoid the number of comparison in MCGP method. Second, the model built by MSGP with linear goal programming indicates that 768 times of simulation can be finished in just one run and still get the appropriate thinning schedule to maximize the carbon sequestration. Finally, the 3-level MCGP and LP are applied to build a model for artificial Sugi. age-classes in Hsinchu Forest District to plan an appropriate thinning schedule for multiple forest management when carbon reduction is considered. The results show that these two techniques can reduce many simulations and comparisons in various thinning regimes and combinations of target goal values to plan an appropriate thinning schedule to achieve the expected carbon sequestration, jobs and soil protection goals.

並列關鍵字

forest management ； multi-choice goal programming ； forest thinning ； carbon sequestration ； multi-segment goal programming

參考文獻

王松永, 張豐丞, 李怡真, 楊賜霖, & 林法勤. (2005). 柳杉疏伐木有效利用之探討. 臺大實驗林研究報告, 19(4), 293-300.

任億安, & 陳宛君. (1996). 六龜試驗林疏伐作業個案採運成本及疏伐效率分析. 台灣林業科學, 11(4), 475-480.

吳俊賢, 林俊成, 李國忠, 陳溢宏, 林麗貞, & 林瑞進. (2005). 森林能源作物之二氧化碳吸存效果與能源產出效率. 台大實驗林研究報告, 19(1), 43-53.

李國忠, 林俊成, 賴建興, & 林麗貞. (2004). 台大實驗林森林生態系不同林分經營策略之碳貯存效果. 臺大實驗林研究報告, 18(4), 261-272.

李國忠, 林俊成, & 陳麗琴. (2000). 台灣杉人工林碳吸存潛力及其成本效益分析. 台灣林業科學, 15(1), 115-123.