Branch and bound is a search method that has proven successful in a variety of optimization problems. A widespread application of branch and bound occurs in integer linear programming problems but the basic idea can be explained in general terms. Suppose one wishes to search a large space of possible solutions to a particular optimization problem. For the sake of definiteness consider a huge collection of points over which is defined an optimality function that measures how optimal each point is as a solution to a particular problem. We wish to find the point with the maximal value of the optimality function. Without any apriori knowledge, one would have to look at each and every point, compute the optimality function for that point, and pick the one that fares the best. Such a brute force procedure is prohibitive as the solution space for all...