Abstract: As an important foundation of materials genome engineering, the establishment of materials database and the study of process parameter optimization have an important impact on materials research and development. Aiming at the current problems that traditional relational databases can not store multi-source high-dimensional heterogeneous alloy data well, and that the materials R&D mode needs to be urgently shifted from the traditional experimental trial-and-error method to a new mode of data-driven materials research, it is proposed to establish an alloy graph database based on ontology and graph data models, and based on this, to search for the optimal combination of process parameters by improving dung-beetle optimization algorithms. The population is initialized using inverse learning, then the position update method of the foraging dung beetle is improved using the variable spiral search strategy to improve the global search capability of the algorithm, and finally the Kersey's variation perturbation and the hybrid strategy are fused at the optimal solution position to generate a new solution to avoid the algorithm from falling into the local optimum. In order to verify the performance of the proposed Improved Dung Beetle Optimization (IDBO) algorithm, the IDBO algorithm is evaluated using 23 classical test functions, and the experimental results show that the IDBO algorithm outperforms other algorithms and significantly improves the convergence speed and optimization accuracy. Meanwhile, the IDBO algorithm is applied to the study of alloy process parameter optimization, and the results show that the IDBO algorithm outperforms other optimization algorithms and obtains the optimal process parameter combinations of alloys, which proves its superiority in process parameter optimization.

Key words: Alloy database, Process parameter optimization, Dung beetle optimization algorithm, Graph data model, Multilayer perception machine