AbstractRaghavachar, has shown the equivalence of zero-one integer programming and a concave quadratic penalty function for a sufficiently large value of the penalty. Kalantari and Rosen found a lower bound for this penalty. It was also shown that this penalty could not be reduced in specific cases. We show that the results generalize to the case where the objective function is any concave function. Equivalent penalty formulation for non-concave functions is also considered.
SubjectsPenalty formulations, Concave minimization, Global optimization
RightsThis Item is protected by copyright and/or related rights.You are free to use this Item in any way that is permitted by the copyright and related rights legislation that applies to your use.For other uses you need to obtain permission from the rights-holder(s).