An interpretable machine-learned model for international oil trade network

Energy security and energy trade are the cornerstones of global economic social development. The structural robustness international oil network (iOTN) plays an important role in economy. We integrate machine learning optimization algorithm, game theory, utility theory for decision-making model that contains benefit endowment cost economies trades. have reconstructed degree, clustering coefficient, closeness iOTN well to verify effectiveness model. In end, policy simulations based on agent-based carried out a more realistic environment. find export-oriented vulnerable being affected than import-oriented after receiving external shocks. Moreover, impact increase decrease friction costs is asymmetrical, there significant differences between organizations.