This paper presents a learning problem from positive examples based on multiple valued minimization paradigm.A new heuristic algorithm for the problem is given.Compared with the existing learning methods from examples
the proposed algorithm is based on deductive reasoning. Furthermore
the algorithm is simple
fast and easy to implement. The comparison of the proposed algorithm with the inductive learning system AQ15 and the Boolean function minimization system SWT is carried out by two illustrations: the knowledge acquisition for the automatic weiding technology design and the Boolean function minimization.