An operational concept capability requirement analysis method based on deep reinforcement learning

• Main Authors: An, Jing, Zhang, Xuechao, Liu, Wei, Rong, Wanting
• Format: Conference Proceeding
• Language: English
• Published: SPIE 21.03.2023
• ISBN: 9781510663459; 1510663452
• ISSN: 0277-786X
• DOI: 10.1117/12.2671678

Based on the perspective of combining qualitative analysis and quantitative calculation, a method of operational concept capability requirement analysis is designed based on deep reinforcement learning. Firstly, it obtains the simulation small sample data set with high reliability based on simulation experiment. Secondly, the operational concept surrogate model is constructed on the empirical data, and the surrogate model is optimized and trained by using the multi-objective optimization algorithm with high credibility simulation data set as input. Finally, the trained surrogate model is interacted with the deep reinforcement learning framework to realize the reverse exploration of operational concept capability requirements.