A Competitive Location Model for the Alternative-fuel Station Allocation Problem
|關鍵字:||競爭性區位模式;替代能源車輛;能源補充站;流量;最大涵蓋問題模式;Competitive location model;Alternative-fuel vehicle;Refueling station;Flow capturing;Maximum coverage problem|
本研究的假設是在一特定區域市場中，已存有競爭對手所設立之替代能源補充站設施，我方也欲在此區域設置替代能源補充站設施。本研究之研究模式是由競爭型流量截取設置模式發展而來，已知需求會透過網路空間經濟地移動到設施接受服務。基於都會區替代能源車輛能源補充站的密度較高，本研究模式中設不考慮部分替代能源車輛補充能源模式之續航力限制假設，透過理想的設施設址，爭取截取(Capturing)到之車流量最大化。針對此問題，本研究除了提出一整數規劃模式(Integer Programming)外，基於求解運算負荷的考量，本研究利用最大涵蓋問題模式(Maximum Coverage Problem)的概念來設計一近似的整數規劃模式，以降低本問題的求解複雜度。透過資料的事先處理，可將原本競爭模式三個索引降低為兩個索引，在求解中、大型問題時能大幅降低解題所需時間。
The consumption of petroleum products is one of the factors causing global warming and serious air pollution. Therefore, many developed countries have devoted great effort to promoting alternative fuel vehicles (AFVs). In addition, due to the nature resource scarcity, the price of traditional fuel is getting higher and higher, some people view the AFV as a good option. However, poor accessibility to refueling stations is one of the major barriers to the adaption of AFVs. However, the cost of establishing the infrastructure to promote the use of AFVs is very high. Thus, the objective of this study is to economically and efficiently determine the location of the AFV refueling facilities. Given a specific market, it is assumed there are some alternative fuel stations set up by an existing competitor, and the new provider plan to set up its own facilities This study is based on competitive flow-capturing location allocation problem (FCLAP) and assume that the drivers/the demand can detour through networks to obtain the service. In the model formulation, we do not take the driving range of AFVs into account, as the density of the facilities in the urban area is relatively high. In order to find the ideal facility locations to maximize the captured traffic flow, we propose an integer programming (IP) model. In addition, with an aim to alleviate the computational load, we make use of the well-known maximum coverage problem to design another IP model to reduce the complexity when solving the problem. By some data pre-processing, we can reduce the indexes of the binary decision variables from three to two, and thus significantly cut the program run time when dealing with the medium or large scale problems. In the numerical experiment, the proposed models and the solution algorithm were tested with a small-sized example network and a middle-sized example network with 7 nodes and 25 nodes respectively. The effects of the number of new facilities in the network and the different facility locations of the existing competitor were examined. In these test problems, we simulate the locations of the existing facilities and the demand flow by random number generators. It is found the solution approach based on the maximum coverage model achieve a better solution quality when compared with the greedy heuristic algorithm in the literature. The test results show the developed model and solution approach can be used for allocating alternative-fuel stations.