A BRANCH AND BOUND ALGORITHM FOR INTEGER AND MIXED-INTEGER LINEAR PROGRAMS.

The algorithm presented is an extension of the Land and Doig branch and bound method combined with the branch selection techniques presented by Beale and Small to solve integer or mixed-integer linear programs. The algorithm obtains the solution by solving a linear program with upper andor lower bounds on selected branch variables. By systematically changing these bounds, and maintaining only the current canonical form, the solution is assured using a minimum of excess computer storage above that required to solve the linear programming problem. Thus the problem can be solved entirely within the computer core, and the problem converges to the solution faster than most other general integer linear programming algorithms. Author