We here present the details of the numerical realization of the recently advanced algorithm developed to identify the fragmentation in heavy ion reactions. This new algorithm is based on the simulated annealing method and is dubbed the simulated annealing clusterization algorithm (SACA). We discuss the different parameters used in the simulated annealing method and present an economical set of the parameters which is based on the extensive analysis carried out for the central and peripheral collisions of Au-Au, Nb-Nb, and Pb-Pb. These parameters are crucial for the success of the algorithm. Our set of optimized parameters gives the same results as the most conservative choice, but is very fast. We also discuss the nucleon and fragment exchange processes which are very important for the energy minimization and finally present the analysis of the reaction dynamics using the new algorithm. This algorithm can be applied whenever one wants to identify which of a given number of constituents form bound objects.