Methods and Algorithms for the Minimization of the Energy Consumed by an Electrical Vehicle

摘要

In recent years, the electrical vehicle energy management has been an important subject and a lot of research has been carried out in this field. However, satisfactory solutions or evident answers for this kind of research do not exist up to now. Indeed, this problem is modeled by a Bang-Bang optimal control problem. The presence of a constraint on a state variable and the Bang-Bang structure of the control make this problem very hard to solve. Recently, an original methodology was developed by Abdelkader Merakeb and Frédéric Messine. They reformulate the initial problem following the construction of a current regulator technique to obtain a global optimization problem which is first discretized and then solved using a Branch and Bound algorithm. This latter has been implemented in MatLab. In this work, we began by providing a compiled code in Fortran 90. The numerical results show that this compiled program performs better than the MATLAB one and the speedups vary, ranging from 2 to 66 depending on the studied instance. After that, we proposed two new heuristic methods to approach the exact bounds used in the Branch and Bound algorithm. These heuristics are very powerful and able to determine the exact solution independently of the studied profile. These significant gains in terms of computation time without loss in quality solution encouraged us to study some extensions of electrical vehicle energy management problems which are the study of long travels and the management of slopes. Numerous numerical tests, showing the efficiency of the proposed approach, are presented.