Students of programming courses are constantly dealing with resolution of lists of problems, aiming to consolidate the learning of programming principles and the development of strategies for general problem resolution. These lists are, most of the times, elaborated without any concern regarding the learning difficulties faced by these students, failing in capturing their individual needs of learning. This paper presents an approach that uses interactive genetic algorithms to dynamically compose lists of problems. With this approach, new problems are suggested to the student considering the challenges in solving the previously suggested problem, adapting the problem to be indicated to these student’s cognitive level.