Using the AMCE Algorithm to High-Efficiently Develop Vehicle Driving Cycles With Road Grade

Abstract

Road grade greatly affects energy consumption and pollutant emission of heavy-duty vehicles. It is therefore necessary to develop representative three-parameter driving cycles while taking road grade into account. However, the low running efficiency of existing methods for developing driving cycles is a serious problem. To improve efficiency, inspired by the idea that intelligent algorithms with self-adaptivity can accelerate convergence, an adaptive Markov chain evolution (AMCE) method is proposed in this study. Based on the characteristics of the evolution strategies of a Markov chain evolution (MCE) satisfying the Markov property, a strategy boundary variable is defined to classify the MCE evolution strategies into two categories, one with the global and one with local search capability. Then, an adaptive probability equation is used to adjust the proportion of the evolution strategies, thus the evolution strategies can satisfy not only the Markov property of driving cycle but is also self-adaptive. By collecting driving data including the elevation information from heavy-duty vehicles driving on a highway, a three-parameter driving cycle with road grade was generated using the proposed method. In comparison with the MCE method, the proposed method can obtain the three-parameter representative driving cycles on the highway, and running efficiency was increased by 43.08% under the given conditions. Additionally, the rationality and necessity of the proposed method are fully verified.

Keywords

• Vehicle driving cycle
• road grade
• self-adaptivity
• Markov chain evolution
• running efficiency