Contributed Session Mon.2.H 2032

Monday, 13:15 - 14:45 h, Room: H 2032

Cluster 11: Integer & mixed-integer programming [...]

Integer programming algorithms II

 

Chair: Serigne Gueye

 

 

Monday, 13:15 - 13:40 h, Room: H 2032, Talk 1

Hilary Paul Williams
The general solution of a mixed integer programme

Coauthor: John N. Hooker

 

Abstract:
We give general formulae for the optimal objective value and solution values of a Mixed Integer Programme (MIP) as a function of the coefficients (objective, matrix and RHS). In order to do this we project out all the variables giving an (attainable) bound on the optimal objective value. This results in the optimal objective value expressed as a finite disjunction of inequalities and linear congruences. While the method is not computationally viable for practical sized MIPs it reveals a lot about their mathematical structure. It also provides a finite method of proving optimality. Hopefully the resultant `dual' solution also provides useful economic interpretations.

 

 

Monday, 13:45 - 14:10 h, Room: H 2032, Talk 2

Serigne Gueye
Using distance variables for the quadratic assignment problem

Coauthor: Philippe Michelon

 

Abstract:
The Quadratic Assignment Problem (QAP) is the problem of allocating facilities to locations such as to minimize the sum of a linear and quadratic cost depending on the distances dkl, between facilities k and l, and flows.
The Minimum Linear Arrangement (MinLA) is a special case of (QAP) where dkl = |k-l|. It has been proposed for MinLA a linear model using distance variables. The lower bound is poor, equal to 0, but can be improved with facets (see [1, 2]). We have independently observed that the distance variables can be used for (QAP), and that some facets are also applicable each time d defines a metric, as the grid graphs in QAPLIB instances. Adding valid inequalities linking D and x, we have tested this model. The result shows a competitive average relative gap of 10.7,% with a minimum of 3.4,%.
{\footnotesize


  1. Liu, W., A. Vannelli. 1995. Generating lower bounds for the linear arrangement problem. Discrete Applied Mathematics 59, p. 137-151.
  2. A. Caprara, A.N. Letchford, and J.J. Salazar-Gonzalez. 2011. Decorous lower
    bounds for minimum linear arrangement. INFORM Journal of Computing, 23(1), p. 26-40.

}

 

  The best way to go for you to know the credible Payday Loans Michigan providers. But at the same time, it acts only with sexual arousal. Viagra has a number of advantages in comparison with injections in the sexual organ or other procedures aimed at treatment of impotency.