What Works? An Anfis-Based Policy Evaluation Framework For Electric Vehicle Technology Development

The application of socio-technical transitions analysis into realms such as future sustainable mobility has revealed the requirement to understand the efficacy of policy measures ex ante. Electrification of the vehicle drivetrain has been promoted as one possible solution to achieve carbon emission and air quality targets. National governments are increasingly forming suites of policy measures to encourage electric vehicle technologies in the transport sector. To evaluate multiple policy measures, a framework based on “adaptive neuro-fuzzy inference systems” (or ANFIS) was developed and is described here. For data generation, the electric vehicle innovation policies of the European Union, United States, Japan, Germany, France and the United Kingdom were analysed and compared with the actual technology development that was measured with patent filings in those regions. The training and validation of the proposed ANFIS framework shows that the model is able to predict the development of electric vehicle technologies in terms of patent filings. The model is subsequently applied to Austria in a predictive capacity to evaluate three proposed policy scenarios. This paper concludes that the developed framework might play a significant role for assisting EV innovation policy-making by enabling ex-ante assessment the effects of different policy-mixes on the technical change.