BUS 336 Lecture Notes - Lecture 10: Integer Programming, Minimax, Normal Distribution
Document Summary
Rounding off the lp solution might not yield the optimal ip solution. Ip objective function value is usually worse (higher cost/lower profit) than the lp value. Ip solutions are usually not at corner points. General integer variables can take any integer value (1,2,3 ) Models containing both integer and general (non integer restricted) decision variables are called mixed ip problems. Modeling binary variables (bv) variables that must take one of two possible values (0,1) Selection problems "control switches" as opposed to "dials" Network models: made entirely of arcs & nodes. Node types: supply nodes, demand nodes, transshipment nodes. Flow balance constraint: net flow per node = (total flow in to node) + ( total flow out of model) This means all out flows will be represented as negative ( ) so that the sumproduct function can apply to all coefficients. Integer values for all constraint coefficients will yield integer decision variable results.