Duality Theory and Optimality Conditions for LPs

Part of the International Series in Operations Research & Management Science book series (ISOR, volume 137)


Associated with every linear programming problem, there is another linear program called its dual, involving a different set of variables, but sharing the same data. When referring to the dual problem of an LP, the original LP is called the primal or the primal problem. Together, the two problems are referred to as a primal, dual pair of linear programs. The names primal, dual for the two problems are coined by Tobias Dantzig, father of George Dantzig, around 1955 in conversations with his son.

A duality type result for systems of linear equations only (no inequalities) is the theorem of alternatives for systems of linear equations (Theorem 1.1 in Sect. 1.2); it has been known for a long time (by the eighteenth century or even earlier), but similar results for systems of linear constraints including linear inequalities were unknown until recently. These important duality-type results for systems of linear constraints including inequalities known as either/or theorems or theorems of alternatives started appearing in published literature beginning in mid-nineteenth century.


Feasible Solution Dual Variable Complementary Pair Sufficient Optimality Condition Complementary Slackness 
These keywords were added by machine and not by the authors. This process is experimental and the keywords may be updated as the learning algorithm improves.


  1. Ahuja RK (1997) The balanced linear programming problem. Eur J Oper Res 101:29–38CrossRefGoogle Scholar
  2. Bazaraa MS, Sherali HD, Shetty CM (2006) Nonlinear programming, theory and algorithms, 3rd edn. Wiley, NYCrossRefGoogle Scholar

Copyright information

© Springer Science+Business Media, LLC 2010

Authors and Affiliations

  1. 1.Dept. Industrial and Operations EngineeringUniversity of MichiganAnn ArborUSA
  2. 2.Systems Engineering DepartmentKing Fahd University of Petroleum and MineralsDhahranSaudi Arabia

Personalised recommendations