19.2 The program in Figure 19.2 initializes the upper bound on the finishing times (bestsofar) to a high value that is obviously a gross overestimate. This makes sure that an optimal schedule is within the bound, and will be found by the program. This policy, although safe, is inefficient because such a loose upper bound does not constrain the search well. Investigate different policies of initializing and changing the upper bound, for example, starting with a very low bound and increasing it if necessary (if no schedule exists within this bound). Compare the run times of various policies. Measure the run time for the case when the bound is immediately set to the true minimal finishing time.