Construction and research of multi-platform knowledge collaboration model based on deep reinforcement learning algorithm

Based on the perspective of knowledge collaboration, this article proposes a multi-platform knowledge collaboration model for managers and learners to make selection decisions. Through data statistics, a deep reinforcement learning algorithm is used to supervise the training data, and a method based on the Markov decision process is proposed to solve the sequence decision optimization problem. Finally, three competition models were constructed by exploring the relationship between collaborative learning motivation, strategy and thinking ability. The structural equation model was used to fit and analyze the three competing models. The results showed that collaborative learning strategies significantly positively affected the five abilities in thinking ability, and also provided managers and learners with the opportunity to conduct research in a "one level and three ends" environment. It provides reference for collaborative learning task design, management and learning.