A sequence-encoded relocation scheme for electric vehicle transport systems

Abstract

This paper designs a resource distribution scheme for city-wide electric vehicle (EV) transport systems, evaluating its performance via a prototype implementation. With the help of computational intelligence and future demand forecasts, the resource distributor tries to enhance the service ratio of EV sharing systems. A genetic algorithm is designed for reasonable response time, focusing on how to encode a relocation schedule so as to represent not just relocation pairs but also operation sequences. The genetic operators are customized for the encoding scheme, while the fitness function estimates relocation distance considering the encoded vector and the number of service men. The experiment result shows that the proposed scheme reduces the resource distribution overhead for the given parameter set and fully benefits from potential operation concurrency, improving the relocation distance by up to 56.9 %, compared with vehicle-by-vehicle moves. © 2014 Springer International Publishing.