Forward and Inverse Models-Based Optimization Method of the Markov Chain to Accurately Design Driving Cycles

The optimization algorithm is crucial for improving the precision and efficiency of driving cycle design. However, when combined with the Markov chain-based driving cycle design method, the effectiveness of the widely used genetic algorithm (GA) may be limited because of constraints in the state transition search space. In this study, a forward and inverse models-based Markov chain evolution (FI-MCE) method is proposed to efficiently and accurately design driving cycles, overcoming the limitations of the GA. Initially, the theoretical relationship between forward and inverse Markov chain models is demonstrated. By incorporating both models, any length of driving cycle can start and end in an idle state, enhancing the efficiency of constructing feasible solution sets. Then, the GA is enhanced with novel mutation and crossover strategies based on these models to simplify operations and overcome limitations. Furthermore, representative driving cycles across different lengths and different geographical regions are designed using FI-MCE to validate its effectiveness. Ultimately, comparing FI-MCE with existing methods demonstrates its superior accuracy and efficiency. Under identical circumstances, FI-MCE consistently outperforms existing methods by swiftly designing more precise driving cycles while ensuring high consistency with database feature distribution. This remarkable capability holds immense potential in overcoming the limitations associated with conventional methodologies.