Thursday, 13:45 - 14:10 h, Room: H 2053

 

John Chinneck
Better placement of local solver launch points for global optimization

Coauthors: Victor Aitken, Laurence Smith

 

Abstract:
NLP solutions are quite sensitive to the launch point provided to the local solver, hence multi-start methods are needed if the global optimum is to be found. The drawback is that local solver launches are expensive. We limit the number of local solver launches by first using very fast approximate methods to explore the variable space to find a small number of promising locations for the local solver launches. We start with a set of random initial points, and then apply the Constraint Consensus (CC) method to quickly move to points that are close to feasibility. Clusters of the CC output points are then automatically identified; these generally correspond to disjoint feasible regions. Finally, the local solver is launched just once from each cluster, greatly improving efficiency. We frequently find a very good solution (if not the optimum solution) with very few local solver launches, and hence in relatively little time. Extensive empirical results are given.

 

Talk 2 of the contributed session Thu.2.H 2053
"Advances in global optimization II" [...]
Cluster 9
"Global optimization" [...]

 

  The best way to go for you to know the credible Michigan Payday Loans providers. In this section we give only a brief summary recommendation for admission of Levitra. Full information can be found in the instructions for receiving medications with vardenafil.