Massive Open Online Courses (MOOC) plays an increasingly important role in changing study habits. MOOC providers o?er thousands of courses with a wide variety of content to the learners. As a result, learners can choose an appropriate learning path based on specifc skills and knowledge. This study presents an alternative normalization method for compromise programming to build an automated learning path recommender system that we published in earlier 2021. In the previous study, we used compromise programming combined with evolutionary algorithms to approach the multi-objective optimization problem. It is effective in guiding the agents in the search process for optimal solutions. However, fnding the optimal solution that closes to the reference point without a suitable normalization and reference point selection method may affect the quality of the solution and make it difcult to assign values to the weight parameters for a particular objective. This paper describes the recommender and alternative editing to the optimization model. We perform experiments with the original data set to compare the effectiveness of the methods. © 2021 Copyright held by the owner/author(s). Publication rights licensed to ACM.