The Haplotyping with Pedigrees (HwP) benchmark suite contains problem instances from the HwP problem. HwP is a variant of the haplotype inference problem, being characterized by two cost functions and preference is given to one of them. Therefore, HwP is a multi-objective combinatorial optimization problem where the optimality criterion is lexicographic. The models used for creating the HwP instances are described in the following papers:

[1] A. Graça, I. Lynce, J. Marques-Silva, and A. Oliveira, "Efficient and Accurate Haplotype Inference by Combining Parsimony and Pedigree Information", ANB'10, volume 6479 of LNCS, 2010.

[2] A. Graça, I. Lynce, J. Marques-Silva, and A. Oliveira, "Haplotype Inference Combining Pedigrees and Unrelated Individuals", WCB'09, 2009.

The HwP benchmark suite includes 100 of the most difficult instances, out of a universe of 945 problem instances used in [2].