网约车合乘出行可有效提高车辆运输效率,与常规网约车出行相比具有显著的碳减排潜力。然而,现实中网约车合乘出行能否真正减少碳排放受多方面因素影响,往往存在较大差异与不确定性。为识别碳减排潜力较大的网约车合乘订单,提出一种基于...网约车合乘出行可有效提高车辆运输效率,与常规网约车出行相比具有显著的碳减排潜力。然而,现实中网约车合乘出行能否真正减少碳排放受多方面因素影响,往往存在较大差异与不确定性。为识别碳减排潜力较大的网约车合乘订单,提出一种基于机器学习的网约车合乘出行碳减排状态预测模型,并解析其碳减排机理。首先,基于成都市真实的网约车合乘订单与轨迹数据,应用COPERT(COmputer Program to calculate Emissions from Road Transport)排放模型分别计算合乘出行碳排放量及其替代的独乘出行碳排放量,进而得到合乘出行相比独乘出行的碳减排量。然后,基于历史的合乘行程碳减排及其订单特征数据,训练XGBoost(eXtreme Gradient Boosting)模型以预测未来潜在合乘出行的碳减排状态。最后,采用ALE(Accumulated Local Effects)分析方法对预测模型进行特征变量解析,以识别影响合乘出行碳减排状态的关键因素。结果显示:研究区域内平均每次网约车合乘出行可减少碳排放307.23 g,但仍有15%的网约车合乘行程未能实现减碳;XGBoost模型可以有效预测网约车合乘出行的碳减排状态,并识别出绕路率、合乘数、重叠率是决定网约车合乘出行碳减排状态的三大关键指标。研究结论可为网约车平台优化合乘订单匹配算法提供理论依据,以实现更高效、更低碳的合乘出行,进一步提高网约车合乘的环境效益。展开更多
Given the growing awareness of the likely catastrophic impacts of climate change and close association of climate change with global emissions of greenhouse gases （of which carbon dioxide is more prominent） , consid...Given the growing awareness of the likely catastrophic impacts of climate change and close association of climate change with global emissions of greenhouse gases （of which carbon dioxide is more prominent） , considerable research efforts have been devoted to the analysis of carbon dioxide （CO2） emissions and its relationship to sustainable development. Now GHG reduction programs have been coming into effect in many developed coun- tries that have more responsibility for historical CO2 emissions, and the studies have become mature. First, the GHG emissions accounting system is more all-inclusive and the methods are more rational with the effort of IPCC from 1995 and all other research- ers related. Second, the responsibility allocation is more rational and fair. Along with the clarity of ＂carbon transfer＂ and ＂carbon leakage＂, the perspective and methodology for allocating regional COz emissions responsibility is turning from production base to consumption base. Third, the decomposition method has become more mature and more complex. For example, the decomposition formulas are including KAYA expression and input-output expres- sion and the decomposition techniques are developed from index analysis to simple average divisia and then adaptive-weighting divisia. Fourth, projection models have become more integrated and long-term. The top-down model and bottom-up model are both inter-embedded and synergetic. Trends above give some advice for the research on CO2 in China, such as emissions factors database construction, deeper-going research on emissions responsibility and structure analysis, promotion of modeling technology and technology-environment database.展开更多
文摘网约车合乘出行可有效提高车辆运输效率,与常规网约车出行相比具有显著的碳减排潜力。然而,现实中网约车合乘出行能否真正减少碳排放受多方面因素影响,往往存在较大差异与不确定性。为识别碳减排潜力较大的网约车合乘订单,提出一种基于机器学习的网约车合乘出行碳减排状态预测模型,并解析其碳减排机理。首先,基于成都市真实的网约车合乘订单与轨迹数据,应用COPERT(COmputer Program to calculate Emissions from Road Transport)排放模型分别计算合乘出行碳排放量及其替代的独乘出行碳排放量,进而得到合乘出行相比独乘出行的碳减排量。然后,基于历史的合乘行程碳减排及其订单特征数据,训练XGBoost(eXtreme Gradient Boosting)模型以预测未来潜在合乘出行的碳减排状态。最后,采用ALE(Accumulated Local Effects)分析方法对预测模型进行特征变量解析,以识别影响合乘出行碳减排状态的关键因素。结果显示:研究区域内平均每次网约车合乘出行可减少碳排放307.23 g,但仍有15%的网约车合乘行程未能实现减碳;XGBoost模型可以有效预测网约车合乘出行的碳减排状态,并识别出绕路率、合乘数、重叠率是决定网约车合乘出行碳减排状态的三大关键指标。研究结论可为网约车平台优化合乘订单匹配算法提供理论依据,以实现更高效、更低碳的合乘出行,进一步提高网约车合乘的环境效益。
基金the helpful funding from the Ministry for Science and technology of China (GrantNo. 2007BAC03A11-04)National Natural Science Foundation of China (Grant No. 41101118)+1 种基金China Postdoctor Science Foundation (Grant No. 20100480438)National Project 973 (Grant No.2012CB95570002)
文摘Given the growing awareness of the likely catastrophic impacts of climate change and close association of climate change with global emissions of greenhouse gases （of which carbon dioxide is more prominent） , considerable research efforts have been devoted to the analysis of carbon dioxide （CO2） emissions and its relationship to sustainable development. Now GHG reduction programs have been coming into effect in many developed coun- tries that have more responsibility for historical CO2 emissions, and the studies have become mature. First, the GHG emissions accounting system is more all-inclusive and the methods are more rational with the effort of IPCC from 1995 and all other research- ers related. Second, the responsibility allocation is more rational and fair. Along with the clarity of ＂carbon transfer＂ and ＂carbon leakage＂, the perspective and methodology for allocating regional COz emissions responsibility is turning from production base to consumption base. Third, the decomposition method has become more mature and more complex. For example, the decomposition formulas are including KAYA expression and input-output expres- sion and the decomposition techniques are developed from index analysis to simple average divisia and then adaptive-weighting divisia. Fourth, projection models have become more integrated and long-term. The top-down model and bottom-up model are both inter-embedded and synergetic. Trends above give some advice for the research on CO2 in China, such as emissions factors database construction, deeper-going research on emissions responsibility and structure analysis, promotion of modeling technology and technology-environment database.
