TDCR: Transformer based decision conflict resolution model for collaborative scheduling

With the strong adaptability and iterative capabilities of artificial intelligent algorithms, more and more enterprises are adopting human–machine collaboration to replace traditional enterprise management methods. So in recent years, the use of evolutionary algorithms and reinforcement learning methods to solve multi-objective optimization has made many breakthroughs, but in the intelligent manufacturing industry, such as production scheduling and inventory management and other aspects of decision optimization, because of the actual situation is too complicated, it is not conducive to determine the appropriate constraints, and then affect the performance of the model. This paper proposes a multi-decision optimization model for collaborative scheduling, which comprises two key components: a conflict resolution strategy module and a decision making module. This paper uses the attention mechanism to generate decision preference vectors in different manufacturing scenarios, so that the conflict resolution strategy can be dynamically changed and adds the keyword mask method close to downstream tasks during training to further improve the performance of the model. Finally, we evaluate the performance of our model in the conflict resolution task by selecting multiple data sets from multiple public data sets, and show satisfactory performance in this task, showing robustness in different scenarios. This study provides a valuable reference for conflict resolution between decision making.