To prevent local optima, a submemeplex is
constructed in each memeplex, which consists of frogs
chosen on the basis of their respective fitness. The
better the fitness, the easier it is chosen. Suppose that
the global optimal solution (position of a frog) is
represented by PX ; and for each submemeplex, PB
and PW respectively represent the best and the worst
solution. The local exploration is implemented in each
submemeplex, i.e., the worst performance frog in the
submemeplex is updated according to the following
modification rule: