Construction of slope-included energy consumption-representative driving cycle for heavy-duty commercial vehicles using multi-dimensional index parameter selection and adaptive crayfish-genetic algorithm

• Novel framework constructs energy-representative driving cycles with road slope. • New Multi-dimensional Index Parameter Selection method identifies core parameters. • Adaptive Crayfish-Genetic Algorithm (ACGA) boosts optimization efficiency by 23.3%. • Achieves only 1.46% fuel consumption deviation, outperforming benchmark methods. • Provides a high-fidelity digital benchmark for energy evaluation and policy-making. Heavy-duty commercial vehicles (HDCVs) exhibit significant deviations between certified fuel consumption and actual operating energy consumption, as standard driving cycles fail to reproduce complex real-road characteristics (e.g., road slopes), which cannot adequately support energy consumption optimization under real road driving cycles. To address this issue, this study proposes an energy consumption-representative driving cycle construction method based on multi-dimensional index parameter selection (MIPS) and an adaptive crayfish-genetic algorithm (ACGA). First, a two-dimensional Markov state transition model integrating vehicle speed and slope is established using the Markov Chain (MC) method to accurately characterize the real driving dynamic characteristics. Second, through the MIPS method, 10 core parameters are optimally selected from 27 candidate characteristics to ensure high-fidelity description of energy consumption laws. Furthermore, aiming at the shortcomings of traditional genetic algorithms (GA), an ACGA with adaptive phase guidance is designed, based on which the MC-ACGA driving cycle synthesis method is constructed and verified through simulation. The results show that the fuel consumption deviation between the driving cycle synthesized by MC-ACGA and the real driving conditions is 1.46%, with accuracy significantly higher than that of traditional MC (5.12%) and MC-GA (2.08%) methods; the computation time is reduced by 65.4% compared with MC and 10.7% compared with MC-GA. The fuel consumption generalization verification deviation under national highway driving cycles is 2.48%, proving the robustness of the method. This study provides a reliable driving cycle benchmark and theoretical support for the accurate evaluation of energy consumption, verification of energy-saving technologies, and formulation of low-carbon policies for HDCVs.