Construction and Empirical Research of a Personalized Learning Path Recommendation Model for “Pharmaceutical E-Commerce” -- A Multi-Objective Optimization Approach Based on Knowledge Graph and Improved Genetic Algorithm

Aiming at the teaching challenges of "Pharmaceutical E-Commerce" such as interdisciplinary integration, complex knowledge systems, and significant differences in students' prior knowledge, this study proposes a Personalized Learning Path Recommendation Model (PLP-RM) integrating Knowledge Graph (KG) and Improved Non-dominated Sorting Genetic Algorithm Ⅱ (INSGA-Ⅱ). Based on curriculum syllabuses, textbooks, and pharmaceutical regulatory policies, a three-dimensional domain knowledge graph of "Concept-Skill-Regulation" is constructed. Combined with Item Response Theory (IRT) to quantify learners' dynamic knowledge states, a personalized learning sequence balancing minimization of cognitive load, maximization of knowledge mastery, and consolidation of weak points is generated through the multi-objective optimization algorithm. A randomized controlled experiment was conducted for 8 weeks with 120 students majoring in Computer Applications at Guangdong Food and Drug Vocational College. The results show that the average post-test score (82.3±4.5), knowledge mastery growth rate (38.2%), and learning satisfaction score (4.23±0.31) of the experimental group are significantly better than those of the control group (73.6±5.2, 21.5%, 3.12±0.45), with all differences achieving statistical significance (p