Research on Path Planning of Computer Fraud and Security Teacher Quality Improvement based on Deep Reinforcement Learning

With the rapid development and popularization of technology, the demand for path planning is also increasing. This paper explores how to optimize and plan the professional development path of teachers in computer fraud prevention education and network security teaching using deep reinforcement learning (DRL) technology. As cybercrime techniques become increasingly complex, higher demands are placed on teachers' knowledge updates and skill development. To solve the convergence speed problem caused by the complex network structure and fixed reward value of the classic deep reinforcement learning algorithms, this paper proposes an improved DQN path planning algorithm and a PER-ERD3QN path planning algorithm. The improved DQN path planning algorithm first embeds a multi-layer perceptron (MLP) into DQN, accelerating the iterative update of the neural network; secondly, this paper introduces a cubic function as a dynamic reward function, returning corresponding reward values based on the state of the object (i.e., its position) after taking an action; finally, it sets an adaptive action selection mechanism, so that the object can rely more on the reward function to choose actions in the later stage of path planning, thus avoiding the probability of randomly choosing actions being too high and missing better choices. For the problem of slow network update speed caused by excessive Q value estimation and insufficient efficient data utilization in the experience pool of the ERDQN path planning algorithm, the PER-ERD3QN path planning algorithm first combines ERDQN with dual Q networks and competitive networks to solve the problem of excessive Q value estimation; secondly, it introduces a priority experience replay mechanism to better utilize the high-quality data in the experience pool. The experimental results show that the PER-ERD3QN path planning algorithm has higher levels in average scores and average reward values, so its network update speed has also been significantly improved. This research not only helps to enhance teachers' professional knowledge and technical application ability in the field of network security, but also provides scientific methodological support for educational institutions to ensure that they can keep up with changing security threats. Furthermore, this study emphasizes the importance of interdisciplinary cooperation and encourages experts from computer science, education, and psychology to work.