Reducing the number of simulations in operation strategy optimization for hybrid electric vehicles

Abstract

The fuel consumption of a simulation model of a real Hybrid Electric Vehicle is optimized on a standardized driving cycle using metaheuristics (PSO, ES, GA). Search space discretization and metamodels are considered for reducing the number of required, time-expensive simulations. Two hybrid metaheuristics for combining the discussed methods are presented. In experiments it is shown that the use of hybrid metaheuristics with discretization and metamodels can lower the number of required simulations without significant loss in solution quality.