A Binary Tuple Encoded Genetic Algorithm for Electric Vehicle Charging Scheduling Optimization

초록

This paper devises a new electric vehicle charging scheduling optimization problem (EVCSOP) by considering the practical property of charging that the charging speed is not consistent during the charging process. The objective of this EVCSOP is to find an optimal charging schedule under given a specific route for severing customers, so as to minimize the overall serving time, which includes the serving time for customers and the charging time for EV. Particularly, this problem is a mixed variable optimization problem involving the discrete variables to determine where to charge and the continuous variables to determine how much electricity to be charged. To effectively solve such a problem, this paper devises a binary tuple encoded genetic algorithm (BTGA). Specifically, it adopts a binary tuple consisting of a Boolean value indicating whether EV needs to charge at the position of one customer and a real value indicating the remaining electricity of EV to encode each dimension. Then, new crossover and mutation operators are designed for GA to evolve the population based on the encoding scheme. Simulation experiments are conducted on the known optimal routes of 8 travelling salesman problem instances with different consumption rates of electricity to compare BTGA with the greedy strategy. Experimental results proves that BTGA noticeably outperforms the greedy method.

키워드