Shift strategy dynamic-optimization method based on deep reinforcement learning

Abstract

The invention belongs to the fields of engineering machinery and vehicle engineering, and particularly relates to a shift strategy dynamic-optimization method based on deep reinforcement learning. Theshift strategy dynamic-optimization method based on deep reinforcement learning comprises the following steps that (1), the state input variable and action output variable of a shift strategy are determined; (2), the Markov decision process of the shift strategy is determined according to the state input variable and the action output variable; (3), according to the shift strategy objective, a reward function of the reinforcement learning shift strategy is established; (4), the deep reinforcement learning shift strategy is solved according to the Markov decision process and the reward function; (5) a prediction Q network calculated in the step (4) is put in a shift strategy controller, during the driving process of engineering machinery and vehicles, the engineering machinery and the vehicles select gears according to the shift strategy controller; and (6), the prediction Q network is regularly updated in the driving process. The shift strategy is updated through a deep reinforcementlearning method to realize the dynamic optimization of the shift strategy.

Description

technical field [0001] The invention belongs to the field of construction machinery and vehicle engineering, and in particular relates to a dynamic optimization method for a shift strategy based on deep reinforcement learning. Background technique [0002] Gear shift strategy is one of the core technologies of construction machinery and vehicle control technology at present, which refers to the law that the gear position changes with the selected parameters during the driving process of construction machinery and vehicles. The solution method is the key consideration in establishing the shift strategy. The solving methods of shift strategy include graphical method, analytical method, genetic algorithm, dynamic programming method and so on. The solution and optimization of the shift strategy is the core direction of the research on the shift strategy, especially the dynamic optimization of the shift strategy is one of the difficulties in the research of the shift strategy. ...

AI Technical Summary

Problems solved by technology

Its disadvantage is that when dynamic programming solves the shifting law, it needs to build a complex state diagram, which is expressed in the form of a table

[0008] In general, most of the existing shift strategy solving or optimization methods cannot be dynamically optimized for the actual driving conditions, and the adaptive ability is poor

Some shift strategies that can be dynamically optimized need to manually formulate dynamic update rules for the shift strategy, which is less intelligent and versatile

Images