基於營運者規劃之無樁式共享電動機車多目標調度規劃模式研究

Abstract

近年來，共享電動機車服務在世界各地迅速發展為一種新興的運輸工具。絕大多數的共享電動機車服務採用無樁式系統，由於無樁式系統允許使用者在該服務的營運範圍內隨時隨地租還運具，使其服務營運者難以按使用者需求適時平衡營運範圍內的機車供給分布。為了解決此種共享運具供需分布不平衡的問題，過去研究針對無樁式共享自行車和共享汽車服務分別發展出多種調度策略，然因這些調度策略未考慮共享電動機車服務的特性，以致共享電動機車服務營運者無法直接採用其策略以執行共享電動機車調度工作；再者，營運者於調度作業規劃期間會面臨不確定性因素，並希望最佳化如促進利潤和滿足使用者需求等多目標衝突問題。基於前述理由，本研究發展出多目標灰色整數規劃模式以解決考慮多部貨車、多場站及不同電量之換電型與充電型機車的電動共享機車調度問題。本研究模式根據三個目標─最大化利潤、最大化共享電動機車服務滿意度和最小化換電工作成本，產生讓多部貨車自多場站出發執行調度機車作業的最適調度路線。此外，本研究提出一種結合灰色整數規劃和ε-constraint方法的模式求解方法，並使用真實的共享機車服務租賃數據進行實例分析工作，將模式應用於臺北市內湖區以處理現實世界複雜的共享機車調度問題；另為反映與日俱增的共享機車服務使用率，本研究對調度問題進行模式的情境分析，結果顯示營運者在規劃與調整調度方式時，應考慮不同目標之間的衝突以及調度後利潤和服務滿意度的變動。本研究模式不僅提出一個結合多種電池維護管理的新穎無樁式共享電動機車調度策略，使用的灰色參數和ε-constraint求解方法更能在不需營運者給予任何先驗知識與偏好，並且有不確定性的情況下，產生多種調度方案。本研究所提出之模型於實務上改善過去方法導致營運者決策的困難，亦有助因應不同狀況作出相應且適當的調度決策。

Electric scooter-sharing services (ESSSs) are innovative mobility solutions that have been expanding rapidly worldwide over the past few years. Most ESSSs are free-floating systems that allow users to rent and return scooters everywhere within their operating areas, so ESSS operators experience difficulty balancing scooter supply and demand spatially and temporally. Multiple relocation strategies have been developed to overcome this imbalance issue in free-floating bike-sharing and car-sharing services. However, ESSS operators may not apply strategies that do not incorporate the characteristics of an ESSS into relocations. Moreover, operators face uncertainties while planning relocations and intend to optimize conflicting objectives, such as profit enhancement and demand fulfillment, which entail trade-offs. For these reasons, this study develops a multiple-objective grey integer programming model to address an ESSS relocation problem with multiple trucks, multiple depots, and multiple commodities (battery-swapping and battery-charging scooters) with different battery levels. In accordance with three optimization objectives, namely, maximizing profits, maximizing service satisfaction, and minimizing battery-swapping costs, the model generates suitable relocation routes for employed trucks starting from their depots to redistribute scooters. Furthermore, a solving approach that combines grey integer programming with the ε-constraint method is proposed to solve the model. Then, authentic ESSS rental data are used in a case study, and the model is applied to Neihu District, Taipei City, to deal with a complicated ESSS relocation problem in the real world. The model is also tested through a scenario analysis to reflect the increasing usage of ESSSs. Results suggest that ESSS operators should consider the conflicts among multiple objectives and the variations in service profits and satisfaction when planning and altering relocation operations. The presented model offers an innovative free-floating ESSS relocation strategy integrated with various battery management schemes, and generates multiple alternatives by using grey parameters and the ε-constraint method without any prior knowledge or preference from an ESSS operator under uncertainty. The model is a practical improvement of the previous methods, which make decision-making difficult for operators, and it helps ESSS operators make appropriate relocation decisions for different scenarios.