Technical Note—An Efficient Approach to Some Cases of Coefficient Variations in Linear Programs

In some cases, a series of linear programs are run in sequence where a set of coefficients in the constraint matrix or in the right hand side is varied in such a way that inequality constraints are relaxed from one run to the next. For this type of variation we present an algorithm that determines in at most K pivots a new feasible solution, if K coefficients are varied. This algorithm relies on post optimality options of linear programming packages. In addition the algorithm proved to be advantageous in determining optimal solutions to the modified problems.