An optimized nonlinear time-varying grey Bernoulli model and its application in forecasting the stock and sales of electric vehicles

An accurate prediction of electric vehicles stock and sales is a prerequisite for planning industrial policies for renewable sources to be used by a transportation system. We propose a novel time-varying grey Bernoulli model to investigate the nonlinear, complexity, and time-varying characteristics associated with electric vehicles stock and sales. We first design the time-varying parameters and a power exponent to explore the nonlinear developing trends of sequences. Subsequently, the cuckoo search algorithm determines optimum solutions because of its competence in dealing with complex optimization problems. Furthermore, its relationship with existing grey prediction models is presented, which demonstrates the flexibility and practicality of the newly-designed model. In order to validate this new model, the global electric vehicles stock and electric vehicles sales in France are predicted in comparison with six benchmark models. As demonstrated by the empirical findings, the proposed model is superior in terms of its capacity for forecasting, confirming its significant potential as a promising tool for electric vehicles stock and sales prediction.