Homogeneous clusters allocation of computational resources

This paper aims to propose a new approach of computational resources clustering, where all of them could be placed in groups with enough conditions to supply some processing demand. This format allows the process result groups to have an equivalent computational power to each other. To allow this, each resource must be evaluated not only by one feature or component, but by a several process relevant elements. This computational power calculation use the concept applied in perceptron neural networks, who use a combination of processed values as a input of a neuron to compile the final value. In a second step, the resources are processed by a farthest hierarchical algorithm, who is in charge to group the elements and balance the total power at the end of the process.