The rapid growth of knowledge and scientific challenges required lifelong continuous education in computational science and engineering.
Computer numerical system representation and computer arithmetic are the basis of numerical computing of scientific models. In this work an adapted student centered and problem based learning strategy is presented. Development of problem solving, effective self directed reasoning and communication skills are promoted. A pilot study was conducted to determine the validity of the proposed alternatives. The study aimed to evaluate the performance of students to solve new problems and effectively describe the problems, the theoretical context and the possible solutions. Preliminary results are presented for a particular population from which the sample is actually drawn.