Optimal driving based trip planning of electric vehicles using evolutionary algorithms: A driving assistance system

Abstract

The existing driving assistance systems (DAS) are not capable to manage the electric vehicle (EV) problems namely insufficiency of charging stations and inadequate range. A novel DAS is presented here to extend the range and overcome other EV drawbacks by suggesting the driver an optimal driving strategy (ODS) continuously throughout trip performing. ODS is decided by solving a multi-objective optimization problem (MOOP), subsequently adopting a multi-criterion decision making technique. Implementation of the DAS in real application requires both better optimization results and low computational time. A study was carried out to investigate the DAS performance with four contending evolutionary algorithms (EAs), NSGAII (a non-dominated sorting multi-objective genetic algorithm), PESA (Pareto envelope-based selection algorithm), PAES (Pareto archived evolution strategy), and SPEA 2 (Strength Pareto evolutionary algorithm). After an initial investigation of EA performances based on different matrices, NSGAII and PESA were found to be most suitable. The natures of decision variables in the Pareto-optimal solutions were analyzed. After an extensive analysis based on different micro-trip structures, it was found that without considering the computational time, PESA solutions possess better convergence and diversity properties than NSGAII solutions. Various approaches were adopted to minimize DAS computation time considering both NSGAII and PESA without significantly compromising the solution’s optimality.