Performance Analysis of an Orienteering Problem-based Trip Scheduler for Electric Vehicles

This paper designs a tour recommendation scheme for electric vehicles and analyzes its performance. Aiming at reducing the time waste brought by long charging time, the recommender not just decides a visiting and charging schedule for the given set of userselected tour spots but also integrate additional places, which have chargers and provide tour activities. Genetic operators, such as reproduction, selection, and mutation, are tailored to create a combinatory schedule of essential selected and optional recommended spots based on a hybrid orienteering model. Its encoding scheme represents a visiting order by a fixedlength integer-valued vector, allowing the omission of recommendable spots. In addition, the fitness function estimates time waste by tracing the battery charge state based on the interdestination distance and stay time along a given route. The performance measurement result obtained from a prototype implementation discovers that our recommendation service can reduce the time waste by up to 12 % for the practical range of parameter setting, compared with the legacy tour scheduling method.