AI-Driven Dynamic Curriculum Design with a Hybrid of Reinforcement Learning and Evolutionary Algorithms

Abstract

The traditional curriculum design frameworks are rigid, which cannot meet the diversified and dynamic learning needs of students and which will not promote students' engagement, knowledge retention, and academic performance optimally. The proposed framework in this paper is a Dynamic Curriculum Design Framework based on Adaptive Intelligent Learning Environments (AILE), which is a combination of Reinforcement Learning (RL) and Evolutionary Algorithms (EA) to provide adaptive, personalized, and optimized learning pathways. The proposed hybrid framework consists of the following five stages: data acquisition, data preprocessing, learner profiling, RL-based curriculum adaptation, and evolutionary optimization. The RL module uses Q-value optimization to adjust curriculum policies during real-time interaction with the learner, and the EA module uses the genetic algorithm to optimize curriculum structures to prevent local optima stagnation by applying selection, crossover, and mutation operations. Personalization accuracy is further improved by the use of min-max normalization and clustering-based learner profiling. The proposed Hybrid RL-EA Framework was evaluated experimentally with educational datasets collected from intelligent learning platforms and was found to significantly outperform all the baseline systems. The curriculum adaptation accuracy was 95.8%, the engagement rate was 93.4%, the efficiency of retaining the course was 91.7%, the course completion rate was 95.2%, academic improvements were 44.1%, and the learning efficiency was 94.8%. Convergence is achieved in 71 iterations, while reduced configurations require 137 iterations. Ablation studies were performed that demonstrated that all components play a critical role in overall performance. The proposed framework provides a scalable and efficient way forward for next-generation, intelligent educational platforms, with measurably better personalization, engagement, and learning results, compared to traditional e-learning and isolated AI-based curriculum platforms.