The paper presents a new Multilevel Ant Stigmergy Algorithm for minimizing the losses in a universal electro-motor by optimizing the independent geometrical parameters of the rotor and the stator. We have proposed a general approach for translating a multi-parameter problem into a tree representation. To find a solution (a path in the tree) we used a stigmergic system based on an ant colony. The average solution obtained with the Multilevel Ant Stigmergy Algorithm is 25% better than a solution recently found using a genetic algorithm.