Multi-start to overcome premature convergence Premature convergence is a problem that always appears when using PSO, in order to overcome the problem we use multi-start. Multi-start is one of the simplest techniques to compute more than one maximum of a function. In this paper, we use this technique to overcome premature convergence. Multi-start, as its name, as soon as a maximum is detected the algorithm is reinitialized in the search space. However, this approach does not guarantee that the algorithm will not convergence to one of the previously detected maximums. In
order to obtain multiple maximums, we use the deflection technique. The approach is as follows. Let be the original objective function and , be m maximums of . Then, a new function is defined as follows,