To predict and evaluate the promotion of electric taxis, drivers’ willingness of owning and driving electric taxis under different operating scenarios are of great importance. However, as electric taxis have not yet been widely adopted in the Hong Kong region, it is not possible to obtain effective operating data for electric taxis, making it difficult to construct a model for taxi drivers’ willingness to drive electric taxis. This paper addresses the challenge of modeling driver behavior during the preoperational phase of electric taxis, where effective empirical data is scarce. To this end, it proposes the generation of a large number of computer agents reflecting public preferences through the extraction of deep information from limited survey data. First, this paper collected operating data of conventional taxis service and drivers’ behavior pattern for electric taxis through questionnaires. Second, extracted the joint probability distribution information, using Monte Carlo stratified sampling to generate computer multiagent to reflect taxi drivers’ behavioral intentions. Then, a method is proposed to verify the partial preferences of multiagent using the operating data of conventional taxis, in order to verify the accuracy and effectiveness of the whole preferences of the multiagent model. Last but not least, based on the multiagent model validated for effectiveness, a simulation analysis of electric taxis promotion in different scenarios is conducted. Additionally, this paper performed sensitivity analyses examining how technical and economic factors including charging speed, charging station coverage ratio, and government subsidies impact electric taxis promotion effectiveness. The multiagent simulation‐based method proposed in this paper effectively supports the quantitative analysis in the evaluation of social system development.
Wu et al. (Thu,) studied this question.