A nondeterministic, maximally parallel methodology for finding the maximum element in a set of numerical values is presented, suitable for being implemented on P systems. Several algorithms of maximum search are then developed for different types of such systems, namely using priorities, nested membranes and linked transport, and their performances are evaluated accordingly. The proposed solutions are expected to find application inside membrane models devoted to compute algorithmic procedures in which the greatest element in a data set must be found. Dynamic algorithms for DNA sequence alignment are an example of such procedures.
